diff --git a/README_us-EN.md b/README_us-EN.md index 20177d88..fbc77553 100644 --- a/README_us-EN.md +++ b/README_us-EN.md @@ -38,6 +38,7 @@ English | [简体中文](README_zh-CN.md) - Segment Anything Model - [SAM](https://arxiv.org/abs/2304.02643): Universal model for natural image segmentation; - [MobileSAM](https://arxiv.org/abs/2306.14289): Faster version of `SAM`; + - [SAM-Med2D](https://github.com/OpenGVLab/SAM-Med2D): 2D Universal model for medical image segmentation (🤗); - [MedSAM](https://arxiv.org/abs/2304.12306): Universal model for medical image segmentation; - [LVMSAM](https://arxiv.org/abs/2306.11925) - [BUID](https://github.com/CVHub520/X-AnyLabeling/tree/main/assets/examples/buid): Ultrasound breast cancer segmentation model; @@ -65,7 +66,7 @@ English | [简体中文](README_zh-CN.md) ### 1.1 Executable -- Download and run the `GUI` version directly from [Baidu Cloud (Access Code: bkdj)](https://pan.baidu.com/s/1cJeRE2wdiYDy05pb5_JqYQ?pwd=bkdj). +- Download and run the `GUI` version directly from [Release tag-v0.2.1](https://github.com/CVHub520/X-AnyLabeling/releases/tag/v0.2.1). Note: - For MacOS: diff --git a/README_zh-CN.md b/README_zh-CN.md index 46b0a104..9713f5e0 100644 --- a/README_zh-CN.md +++ b/README_zh-CN.md @@ -14,7 +14,7 @@
-[帮助文档](./docs/Q&A.md) +👉[帮助文档](./docs/Q&A.md)👈 简体中文 | [English](README_us-EN.md) @@ -41,6 +41,7 @@ - Segment Anything Model - [SAM](https://arxiv.org/abs/2304.02643): 通用自然图像分割一切模型； - [MobileSAM](https://arxiv.org/abs/2306.14289): 快速版 `SAM`； + - [SAM-Med2D](https://github.com/OpenGVLab/SAM-Med2D): 2D医学图像分割一切模型（🤗推荐）; - [MedSAM](https://arxiv.org/abs/2304.12306): 通用医学图像分割一切模型； - [LVMSAM](https://arxiv.org/abs/2306.11925) - [BUID](https://github.com/CVHub520/X-AnyLabeling/tree/main/assets/examples/buid): 超声乳腺癌分割模型； @@ -69,7 +70,7 @@ ### 1.1 可执行文件 -- 从[百度网盘(提取码: bkdj)](https://pan.baidu.com/s/1cJeRE2wdiYDy05pb5_JqYQ?pwd=bkdj)下载并运行`GUI`版本直接运行。 +- 从[百度网盘(提取码: cn0c)](https://pan.baidu.com/s/1HeEoVc-_DI5EuSU3MipqJQ?pwd=cn0c)下载并运行`GUI`版本直接运行。 注意： - 对于MacOS： diff --git a/anylabeling/app_info.py b/anylabeling/app_info.py index 2c5dfd6c..376916d9 100644 --- a/anylabeling/app_info.py +++ b/anylabeling/app_info.py @@ -1,4 +1,4 @@ __appname__ = "X-AnyLabeling" __appdescription__ = "Advanced Auto Labeling Solution with Added Features" -__version__ = "0.2.0" +__version__ = "0.2.1" __preferred_device__ = "CPU" # GPU or CPU diff --git a/anylabeling/configs/auto_labeling/models.yaml b/anylabeling/configs/auto_labeling/models.yaml index 82c96c0b..58cb8982 100644 --- a/anylabeling/configs/auto_labeling/models.yaml +++ b/anylabeling/configs/auto_labeling/models.yaml @@ -1,3 +1,5 @@ +- model_name: "sam_med2d_vit_b-r20230901" + config_file: ":/sam_med2d_vit_b.yaml" - model_name: "segment_anything_vit_b_quant-r20230810" config_file: ":/segment_anything_vit_b_quant.yaml" - model_name: "segment_anything_vit_b-r20230810" @@ -28,6 +30,14 @@ config_file: ":/medsam_vit_b_quant.yaml" - model_name: "mobile_sam_vit_h-r20230810" config_file: ":/mobile_sam_vit_h.yaml" +- model_name: "clrnet_tusimple_r18-r20230901" + config_file: ":/clrnet_tusimple_r18.yaml" +- model_name: "yolox_l_dwpose_ucoco-r20230820" + config_file: ":/yolox_l_dwpose_ucoco.yaml" +- model_name: "rtdetr_r50-r20230520" + config_file: ":/rtdetr_r50.yaml" +- model_name: "yolov5s_resnet50-r20230520" + config_file: ":/yolov5s_resnet50.yaml" - model_name: "yolov5n-r20230520" config_file: ":/yolov5n.yaml" - model_name: "yolov5s-r20230520" @@ -81,12 +91,4 @@ - model_name: "yolo-nas-m-r20230615" config_file: ":/yolo_nas_m.yaml" - model_name: "yolo-nas-l-r20230615" - config_file: ":/yolo_nas_l.yaml" -- model_name: "yolov5s_resnet50-r20230520" - config_file: ":/yolov5s_resnet50.yaml" -- model_name: "rtdetr_r50-r20230520" - config_file: ":/rtdetr_r50.yaml" -- model_name: "yolox_l_dwpose_ucoco-r20230820" - config_file: ":/yolox_l_dwpose_ucoco.yaml" -- model_name: "clrnet_tusimple_r18-r20230901" - config_file: ":/clrnet_tusimple_r18.yaml" \ No newline at end of file + config_file: ":/yolo_nas_l.yaml" \ No newline at end of file diff --git a/anylabeling/configs/auto_labeling/sam_med2d_vit_b.yaml b/anylabeling/configs/auto_labeling/sam_med2d_vit_b.yaml new file mode 100644 index 00000000..97cd545b --- /dev/null +++ b/anylabeling/configs/auto_labeling/sam_med2d_vit_b.yaml @@ -0,0 +1,6 @@ +type: sam_med2d +name: sam-med2d_vit_b-r20230901 +display_name: SAM-Med2D-256x (ViT-Base) +encoder_model_path: https://github.com/CVHub520/X-AnyLabeling/releases/download/v0.2.0/sam-med2d_b.encoder.onnx +decoder_model_path: https://github.com/CVHub520/X-AnyLabeling/releases/download/v0.2.0/sam-med2d_b.decoder.onnx +input_size: 256 diff --git a/anylabeling/services/auto_labeling/model.py b/anylabeling/services/auto_labeling/model.py index 74e17663..112953b4 100644 --- a/anylabeling/services/auto_labeling/model.py +++ b/anylabeling/services/auto_labeling/model.py @@ -30,7 +30,7 @@ class Model(QObject): BASE_DOWNLOAD_URL = ( - "https://github.com/CVHub520/X-AnyLabeling/releases/tag/v0.2.0" + "https://github.com/CVHub520/X-AnyLabeling/releases/tag/v0.2.1" ) class Meta(QObject): diff --git a/anylabeling/services/auto_labeling/model_manager.py b/anylabeling/services/auto_labeling/model_manager.py index f222be15..6f132c1c 100644 --- a/anylabeling/services/auto_labeling/model_manager.py +++ b/anylabeling/services/auto_labeling/model_manager.py @@ -165,7 +165,8 @@ def load_custom_model(self, config_file): or "name" not in model_config or model_config["type"] not in [ - "segment_anything", + "segment_anything", + "sam_med2d", "yolov5", "yolov6", "yolov7", @@ -455,7 +456,30 @@ def _load_model(self, model_id): ) ) return + # Request next files for prediction + self.request_next_files_requested.emit() + elif model_config["type"] == "sam_med2d": + from .sam_med2d import SAM_Med2D + try: + model_config["model"] = SAM_Med2D( + model_config, on_message=self.new_model_status.emit + ) + self.auto_segmentation_model_selected.emit() + except Exception as e: # noqa + print( + "Error in loading model: {error_message}".format( + error_message=str(e) + ) + ) + self.new_model_status.emit( + self.tr( + "Error in loading model: {error_message}".format( + error_message=str(e) + ) + ) + ) + return # Request next files for prediction self.request_next_files_requested.emit() elif model_config["type"] == "yolov5_cls": @@ -580,7 +604,7 @@ def set_auto_labeling_marks(self, marks): """ if ( self.loaded_model_config is None - or self.loaded_model_config["type"] != "segment_anything" + or self.loaded_model_config["type"] not in ["segment_anything", "sam_med2d"] ): return self.loaded_model_config["model"].set_auto_labeling_marks(marks) @@ -666,8 +690,8 @@ def on_next_files_changed(self, next_files): if self.loaded_model_config is None: return - # Currently only segment_anything model supports this feature - if self.loaded_model_config["type"] != "segment_anything": + # Currently only segment_anything-like model supports this feature + if self.loaded_model_config["type"] not in ["segment_anything", "sam_med2d"]: return self.loaded_model_config["model"].on_next_files_changed(next_files) diff --git a/anylabeling/services/auto_labeling/sam_med2d.py b/anylabeling/services/auto_labeling/sam_med2d.py new file mode 100644 index 00000000..c68e6717 --- /dev/null +++ b/anylabeling/services/auto_labeling/sam_med2d.py @@ -0,0 +1,439 @@ +import logging +import os +import traceback + +import cv2 +import numpy as np +import onnxruntime as ort +from copy import deepcopy +from PyQt5 import QtCore +from PyQt5.QtCore import QThread +from PyQt5.QtCore import QCoreApplication + +from anylabeling.utils import GenericWorker +from anylabeling.views.labeling.shape import Shape +from anylabeling.views.labeling.utils.opencv import qt_img_to_rgb_cv_img + +from .lru_cache import LRUCache +from .model import Model +from .types import AutoLabelingResult + +class SegmentAnythingONNX: + + def __init__(self, encoder_model_path, decoder_model_path) -> None: + # Basic hyp-parameters + self.pixel_mean = np.array([123.675, 116.28, 103.53]) + self.pixel_std = np.array([58.395, 57.12, 57.375]) + + # Load models + providers = ort.get_available_providers() + + # Pop TensorRT Runtime due to crashing issues + # TODO: Add back when TensorRT backend is stable + providers = [p for p in providers if p != "TensorrtExecutionProvider"] + + if providers: + logging.info( + "Available providers for ONNXRuntime: %s", ", ".join(providers) + ) + else: + logging.warning("No available providers for ONNXRuntime") + + self.encoder_session = ort.InferenceSession(encoder_model_path, providers=providers) + self.decoder_session = ort.InferenceSession(decoder_model_path, providers=providers) + + self.encoder_input_name = self.encoder_session.get_inputs()[0].name + self.encoder_input_shape = self.encoder_session.get_inputs()[0].shape + self.encoder_input_size = self.encoder_input_shape[-2:] + + def run_encoder(self, encoder_inputs): + """Run encoder""" + output = self.encoder_session.run(None, encoder_inputs) + image_embedding = output[0] + return image_embedding + + def transform(self, input_image: np.ndarray) -> np.ndarray: + """image transform + + This function can convert the input image to the required input format for vit. + + Args: + input_image (np.ndarray): input image, the image type should be RGB. + + Returns: + np.ndarray: transformed image. + """ + # Normalization + input_image = (input_image - self.pixel_mean) / self.pixel_std + + # Resize + input_image = cv2.resize(input_image, self.encoder_input_size, cv2.INTER_NEAREST) + + # HWC -> CHW + input_image = input_image.transpose((2, 0, 1)) + + # CHW -> NCHW + input_image = np.expand_dims(input_image, 0).astype(np.float32) + + return input_image + + def encode(self, cv_image): + """ + Calculate embedding and metadata for a single image. + """ + original_size = cv_image.shape[:2] + encoder_inputs = { + self.encoder_input_name: self.transform(cv_image), + } + + image_embedding = self.run_encoder(encoder_inputs) + return { + "image_embedding": image_embedding, + "original_size": original_size, + } + + def get_input_points(self, prompt): + """Get input points""" + points = [] + labels = [] + for mark in prompt: + if mark["type"] == "point": + points.append(mark["data"]) + labels.append(mark["label"]) + elif mark["type"] == "rectangle": + points.append([mark["data"][0], mark["data"][1]]) # top left + points.append([mark["data"][2], mark["data"][3]]) # bottom right + labels.append(2) + labels.append(3) + points, labels = np.array(points).astype(np.float32), np.array(labels).astype(np.float32) + return points, labels + + def apply_coords(self, coords, original_size, new_size): + old_h, old_w = original_size + new_h, new_w = new_size + coords = deepcopy(coords).astype(float) + coords[..., 0] = coords[..., 0] * (new_w / old_w) + coords[..., 1] = coords[..., 1] * (new_h / old_h) + return coords + + def run_decoder(self, image_embedding, original_size, prompt): + """Run decoder""" + point_coords, point_labels = self.get_input_points(prompt) + + if point_coords is None or point_labels is None: + raise ValueError("Unable to segment, please input at least one box or point.") + + h, w = self.encoder_input_size + if image_embedding.shape != (1, 256, int(h/16), int(w/16)): + raise ValueError("Got wrong embedding shape!") + + has_mask_input = np.zeros(1, dtype=np.float32) + mask_input = np.zeros((1, 1, int(h/4), int(w/4)), dtype=np.float32) + + if point_coords is not None: + if isinstance(point_coords, list): + point_coords = np.array(point_coords, dtype=np.float32) + if isinstance(point_labels, list): + point_labels = np.array(point_labels, dtype=np.float32) + + if point_coords is not None: + point_coords = self.apply_coords(point_coords, original_size, self.encoder_input_size).astype(np.float32) + point_coords = np.expand_dims(point_coords, axis=0) + point_labels = np.expand_dims(point_labels, axis=0) + + assert point_coords.shape[0] == 1 and point_coords.shape[-1] == 2 + assert point_labels.shape[0] == 1 + input_dict = {"image_embeddings": image_embedding, + "point_coords": point_coords, + "point_labels": point_labels, + "mask_input": mask_input, + "has_mask_input": has_mask_input, + "orig_im_size": np.array(original_size, dtype=np.float32)} + masks, _, _ = self.decoder_session.run(None, input_dict) + + return masks + + def predict_masks(self, embedding, prompt): + """ + Predict masks for a single image. + """ + masks = self.run_decoder( + embedding["image_embedding"], + embedding["original_size"], + prompt, + ) + + return masks + +class SAM_Med2D(Model): + """Segmentation model using SAM_Med2D""" + + class Meta: + required_config_names = [ + "type", + "name", + "display_name", + "encoder_model_path", + "decoder_model_path", + ] + widgets = [ + "output_label", + "output_select_combobox", + "button_add_point", + "button_remove_point", + "button_add_rect", + "button_clear", + "button_finish_object", + ] + output_modes = { + "polygon": QCoreApplication.translate("Model", "Polygon"), + "rectangle": QCoreApplication.translate("Model", "Rectangle"), + } + default_output_mode = "polygon" + + def __init__(self, config_path, on_message) -> None: + # Run the parent class's init method + super().__init__(config_path, on_message) + self.input_size = self.config["input_size"] + + # Get encoder and decoder model paths + encoder_model_abs_path = self.get_model_abs_path( + self.config, "encoder_model_path" + ) + if not encoder_model_abs_path or not os.path.isfile( + encoder_model_abs_path + ): + raise FileNotFoundError( + QCoreApplication.translate( + "Model", + "Could not download or initialize encoder of Segment Anything.", + ) + ) + decoder_model_abs_path = self.get_model_abs_path( + self.config, "decoder_model_path" + ) + if not decoder_model_abs_path or not os.path.isfile( + decoder_model_abs_path + ): + raise FileNotFoundError( + QCoreApplication.translate( + "Model", + "Could not download or initialize decoder of Segment Anything.", + ) + ) + + # Load models + self.model = SegmentAnythingONNX( + encoder_model_abs_path, decoder_model_abs_path + ) + + # Mark for auto labeling + # points, rectangles + self.marks = [] + + # Cache for image embedding + self.cache_size = 10 + self.preloaded_size = self.cache_size - 3 + self.image_embedding_cache = LRUCache(self.cache_size) + + # Pre-inference worker + self.pre_inference_thread = None + self.pre_inference_worker = None + self.stop_inference = False + + def set_auto_labeling_marks(self, marks): + """Set auto labeling marks""" + self.marks = marks + + def post_process(self, masks): + """ + Post process masks + """ + # Find contours + masks[masks > 0.0] = 255 + masks[masks <= 0.0] = 0 + masks = masks.astype(np.uint8) + contours, _ = cv2.findContours( + masks, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_NONE + ) + + # Refine contours + approx_contours = [] + for contour in contours: + # Approximate contour + epsilon = 0.001 * cv2.arcLength(contour, True) + approx = cv2.approxPolyDP(contour, epsilon, True) + approx_contours.append(approx) + + # Remove too big contours ( >90% of image size) + if len(approx_contours) > 1: + image_size = masks.shape[0] * masks.shape[1] + areas = [cv2.contourArea(contour) for contour in approx_contours] + filtered_approx_contours = [ + contour + for contour, area in zip(approx_contours, areas) + if area < image_size * 0.9 + ] + + # Remove small contours (area < 20% of average area) + if len(approx_contours) > 1: + areas = [cv2.contourArea(contour) for contour in approx_contours] + avg_area = np.mean(areas) + + filtered_approx_contours = [ + contour + for contour, area in zip(approx_contours, areas) + if area > avg_area * 0.2 + ] + approx_contours = filtered_approx_contours + + # Contours to shapes + shapes = [] + if self.output_mode == "polygon": + for approx in approx_contours: + # Scale points + points = approx.reshape(-1, 2) + points[:, 0] = points[:, 0] + points[:, 1] = points[:, 1] + points = points.tolist() + if len(points) < 3: + continue + points.append(points[0]) + + # Create shape + shape = Shape(flags={}) + for point in points: + point[0] = int(point[0]) + point[1] = int(point[1]) + shape.add_point(QtCore.QPointF(point[0], point[1])) + shape.shape_type = "polygon" + shape.closed = True + shape.fill_color = "#000000" + shape.line_color = "#000000" + shape.line_width = 1 + shape.label = "AUTOLABEL_OBJECT" + shape.selected = False + shapes.append(shape) + elif self.output_mode == "rectangle": + x_min = 100000000 + y_min = 100000000 + x_max = 0 + y_max = 0 + for approx in approx_contours: + # Scale points + points = approx.reshape(-1, 2) + points[:, 0] = points[:, 0] + points[:, 1] = points[:, 1] + points = points.tolist() + if len(points) < 3: + continue + + # Get min/max + for point in points: + x_min = min(x_min, point[0]) + y_min = min(y_min, point[1]) + x_max = max(x_max, point[0]) + y_max = max(y_max, point[1]) + + # Create shape + shape = Shape(flags={}) + shape.add_point(QtCore.QPointF(x_min, y_min)) + shape.add_point(QtCore.QPointF(x_max, y_max)) + shape.shape_type = "rectangle" + shape.closed = True + shape.fill_color = "#000000" + shape.line_color = "#000000" + shape.line_width = 1 + shape.label = "AUTOLABEL_OBJECT" + shape.selected = False + shapes.append(shape) + + return shapes + + def predict_shapes(self, image, filename=None) -> AutoLabelingResult: + """ + Predict shapes from image + """ + + if image is None or not self.marks: + return AutoLabelingResult([], replace=False) + + shapes = [] + try: + # Use cached image embedding if possible + cached_data = self.image_embedding_cache.get(filename) + if cached_data is not None: + image_embedding = cached_data + else: + cv_image = qt_img_to_rgb_cv_img(image, filename) + if self.stop_inference: + return AutoLabelingResult([], replace=False) + image_embedding = self.model.encode(cv_image) + self.image_embedding_cache.put( + filename, + image_embedding, + ) + if self.stop_inference: + return AutoLabelingResult([], replace=False) + masks = self.model.predict_masks(image_embedding, self.marks) + if len(masks.shape) == 4: + masks = masks[0][0] + else: + masks = masks[0] + shapes = self.post_process(masks) + except Exception as e: # noqa + logging.warning("Could not inference model") + logging.warning(e) + traceback.print_exc() + return AutoLabelingResult([], replace=False) + + result = AutoLabelingResult(shapes, replace=False) + return result + + def unload(self): + self.stop_inference = True + if self.pre_inference_thread: + self.pre_inference_thread.quit() + + def preload_worker(self, files): + """ + Preload next files, run inference and cache results + """ + files = files[: self.preloaded_size] + for filename in files: + if self.image_embedding_cache.find(filename): + continue + image = self.load_image_from_filename(filename) + if image is None: + continue + if self.stop_inference: + return + cv_image = qt_img_to_rgb_cv_img(image) + image_embedding = self.model.encode(cv_image) + self.image_embedding_cache.put( + filename, + image_embedding, + ) + + def on_next_files_changed(self, next_files): + """ + Handle next files changed. This function can preload next files + and run inference to save time for user. + """ + if ( + self.pre_inference_thread is None + or not self.pre_inference_thread.isRunning() + ): + self.pre_inference_thread = QThread() + self.pre_inference_worker = GenericWorker( + self.preload_worker, next_files + ) + self.pre_inference_worker.finished.connect( + self.pre_inference_thread.quit + ) + self.pre_inference_worker.moveToThread(self.pre_inference_thread) + self.pre_inference_thread.started.connect( + self.pre_inference_worker.run + ) + self.pre_inference_thread.start() + diff --git a/anylabeling/services/auto_labeling/segment_anything.py b/anylabeling/services/auto_labeling/segment_anything.py index 5eb1337c..1b5f0704 100644 --- a/anylabeling/services/auto_labeling/segment_anything.py +++ b/anylabeling/services/auto_labeling/segment_anything.py @@ -209,9 +209,10 @@ def predict_shapes(self, image, filename=None) -> AutoLabelingResult: """ Predict shapes from image """ + print(f"image={image}, self.marks={self.marks}") if image is None or not self.marks: return AutoLabelingResult([], replace=False) - + shapes = [] try: # Use cached image embedding if possible @@ -289,4 +290,5 @@ def on_next_files_changed(self, next_files): self.pre_inference_thread.started.connect( self.pre_inference_worker.run ) - self.pre_inference_thread.start() \ No newline at end of file + self.pre_inference_thread.start() + diff --git a/assets/examples/lane/lane_demo.json b/assets/examples/lane/lane_demo.json new file mode 100644 index 00000000..6e13c5ad --- /dev/null +++ b/assets/examples/lane/lane_demo.json @@ -0,0 +1,1401 @@ +{ + "version": "0.2.0", + "flags": {}, + "shapes": [ + { + "label": "lane1", + "text": "", + "points": [ + [ + 227.0, + 710.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 236.0, + 700.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 245.0, + 690.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 254.0, + 680.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 264.0, + 670.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 273.0, + 660.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 282.0, + 650.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 291.0, + 640.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 301.0, + 630.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 310.0, + 620.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 319.0, + 610.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 328.0, + 600.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 338.0, + 590.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 347.0, + 580.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 356.0, + 570.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 365.0, + 560.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 375.0, + 550.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 384.0, + 540.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 393.0, + 530.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 403.0, + 520.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 412.0, + 510.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 421.0, + 500.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 430.0, + 490.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 440.0, + 480.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 449.0, + 470.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 458.0, + 459.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 467.0, + 450.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 477.0, + 440.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 486.0, + 430.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 495.0, + 420.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 505.0, + 410.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 514.0, + 400.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 523.0, + 390.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 533.0, + 380.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 542.0, + 369.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 552.0, + 360.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 561.0, + 350.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 570.0, + 340.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 580.0, + 330.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 589.0, + 320.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 599.0, + 310.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 608.0, + 300.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 618.0, + 290.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 627.0, + 280.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 637.0, + 270.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane1", + "text": "", + "points": [ + [ + 646.0, + 260.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane2", + "text": "", + "points": [ + [ + 1252.0, + 420.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane2", + "text": "", + "points": [ + [ + 1218.0, + 410.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane2", + "text": "", + "points": [ + [ + 1185.0, + 400.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane2", + "text": "", + "points": [ + [ + 1151.0, + 390.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane2", + "text": "", + "points": [ + [ + 1117.0, + 380.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane2", + "text": "", + "points": [ + [ + 1084.0, + 369.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane2", + "text": "", + "points": [ + [ + 1050.0, + 360.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane2", + "text": "", + "points": [ + [ + 1016.0, + 350.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane2", + "text": "", + "points": [ + [ + 982.0, + 340.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane2", + "text": "", + "points": [ + [ + 949.0, + 330.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane2", + "text": "", + "points": [ + [ + 915.0, + 320.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane2", + "text": "", + "points": [ + [ + 881.0, + 310.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane2", + "text": "", + "points": [ + [ + 848.0, + 300.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane2", + "text": "", + "points": [ + [ + 814.0, + 290.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane2", + "text": "", + "points": [ + [ + 781.0, + 280.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane2", + "text": "", + "points": [ + [ + 748.0, + 270.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 1268.0, + 700.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 1255.0, + 690.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 1242.0, + 680.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 1228.0, + 670.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 1215.0, + 660.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 1202.0, + 650.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 1188.0, + 640.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 1175.0, + 630.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 1162.0, + 620.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 1148.0, + 610.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 1135.0, + 600.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 1122.0, + 590.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 1108.0, + 580.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 1095.0, + 570.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 1082.0, + 560.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 1068.0, + 550.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 1055.0, + 540.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 1042.0, + 530.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 1028.0, + 520.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 1015.0, + 510.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 1002.0, + 500.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 988.0, + 490.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 975.0, + 480.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 962.0, + 470.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 948.0, + 459.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 935.0, + 450.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 922.0, + 440.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 908.0, + 430.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 895.0, + 420.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 882.0, + 410.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 868.0, + 400.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 855.0, + 390.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 842.0, + 380.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 828.0, + 369.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 815.0, + 360.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 801.0, + 350.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 788.0, + 340.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 774.0, + 330.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 761.0, + 320.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 748.0, + 310.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 734.0, + 300.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 721.0, + 290.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 708.0, + 280.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 695.0, + 270.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + }, + { + "label": "lane3", + "text": "", + "points": [ + [ + 682.0, + 260.0 + ] + ], + "group_id": null, + "shape_type": "point", + "flags": {} + } + ], + "imagePath": "lane_demo.jpg", + "imageData": null, + "imageHeight": 720, + "imageWidth": 1280 +} \ No newline at end of file diff --git a/scripts/zip_models.py b/scripts/zip_models.py index d5fabc4d..d03c4d44 100644 --- a/scripts/zip_models.py +++ b/scripts/zip_models.py @@ -29,7 +29,7 @@ def get_filename_from_url(url): model_config = yaml.load( open(model_config_path + config_file, "r"), Loader=yaml.FullLoader ) - if model_config["type"] == "segment_anything": + if model_config["type"] in ["segment_anything", "sam_med2d"]: download_links.append(model_config["encoder_model_path"]) download_links.append(model_config["decoder_model_path"]) else: @@ -40,7 +40,7 @@ def get_filename_from_url(url): # Save model config # Rewrite model's urls - if model_config["type"] == "segment_anything": + if model_config["type"] in ["segment_anything", "sam_med2d"]: model_config["encoder_model_path"] = get_filename_from_url( model_config["encoder_model_path"] )