网站设计流程是什么,网页如何保存密码,wordpress微信启动,电子商务网站建设与维护论文文章首发见博客#xff1a;https://mwhls.top/4869.html。 无图/格式错误/后续更新请见首发页。 更多更新请到mwhls.top查看 欢迎留言提问或批评建议#xff0c;私信不回。 Github - 开源代码及Readme Blog - 工具介绍 摘要#xff1a;基于TorchCam实现ReID的热力图可视化的… 文章首发见博客https://mwhls.top/4869.html。 无图/格式错误/后续更新请见首发页。 更多更新请到mwhls.top查看 欢迎留言提问或批评建议私信不回。 Github - 开源代码及Readme Blog - 工具介绍 摘要基于TorchCam实现ReID的热力图可视化的工具介绍。 场景
想为科研添砖加瓦因此需要一个好看的可视化。或者只是想了解一下模型的效果。
使用示例
开源地址asdTools/asdTools/Tools/Image/VisualizeHeatmapOfReID.py at main处理结果示例asdTools/Sample/VisualizeHeatmapOfReID at main实例化一个VisualizeHeatmapOfReID对象将 模型、图片文件夹、测试时transform 传入即可。 使用示例见if __name__ __main__: 安装 torch-cam: pip install torchcam or conda install -c frgfm torchcam, GitHub: https://github.com/frgfm/torch-cam from torchvision.transforms.functional import to_pil_imagefrom asdTools.Classes.Image.ImageBase import ImageBasefrom torchcam.utils import overlay_maskfrom torchcam.methods import GradCAMclass VisualizeHeatmapOfReID(ImageBase): Sample: Sample/VisualizeHeatmapOfReID使用torch-cam可视化热力图仅测试于ReID模型。Visualize heatmap by torch-cam, test only on ReID.torch-cam: pip install torchcam or conda install -c frgfm torchcam, GitHub: https://github.com/frgfm/torch-camdef __init__(self, **kwargs) - None:super().__init__(multipleFilesTrue, **kwargs)def __call__(self, model, imgs_dir:str, transform,img_ext:list[png, jpg, jpeg],torchCamMethodGradCAM,devicecuda) - str:self.run(model, imgs_dir, transform, img_ext, torchCamMethod, device)def run(self, model, imgs_dir:str, transform, img_ext:list, torchCamMethod, device) - str:self.begining()# get paths of imgs from img_dirimgs_path self.get_paths_from_dir(imgs_dir)self.log(f{len(imgs_path)} files found in {imgs_dir})imgs_path self.filter_ext(imgs_path, img_ext)self.log(f{len(imgs_path)} images found after filter extension by {img_ext})# init torchCammodel.to(device).eval()cam_extractor torchCamMethod(model)for i, img_path in enumerate(imgs_path):# model(x)img self.read_img(img_path)x transform(img).unsqueeze(0).to(device)out model(x)# visualize featureactivation_map cam_extractor(class_idx0, scoresout.unsqueeze(0))[0]result overlay_mask(img, to_pil_image(activation_map, modeF), alpha0.5)# save imgsave_path self.remove_root_of_path(pathimg_path, rootimgs_dir)save_middle_dir self.get_dir_of_file(save_path)save_name self.get_name_of_file(save_path, True)save_path self.save_image(result, output_middle_dirsave_middle_dir, output_filesave_name)self.log(f{i1}/{len(imgs_path)}: the heatmap of {img_path} has been saved to {save_path}.)self.done()if __name__ __main__: ---------- usage example ---------- -- you can simply insert VisualizeHeatmapOfReID() in the train.py or main.py-- and then set value of imgs_dirfrom asdTools.Tools.Image.VisualizeHeatmapOfReID import VisualizeHeatmapOfReIDfrom models.img_resnet import ResNet50import data.img_transforms as Timgs_dir ../../../../Datasets/PRCCweight logs/0/baseline.pth.tarcheckpoint torch.load(weight)model ResNet50(config)model.load_state_dict(checkpoint[model_state_dict])model.cuda().eval()transform_test T.Compose([T.Resize((256, 128)),T.ToTensor(),T.Normalize(mean[0.485, 0.456, 0.406], std[0.229, 0.224, 0.225]),])color_heatmap VisualizeHeatmapOfReID()color_heatmap(model, imgs_dir, transform_test)
