This repo provides the Graphic User Interface for Deep Learning models to realize "One-button" to use these models. In detail, it provides plugins in ImagePy to run the DL models.
让深度学习算法触手可及、一键调用，不必每次在命令行进行复杂配置。
Download the model folder, and place it in the imagepy/plugins folder.
Then the menu for this model will appear in the imagepy menu bar.
只需将要使用的模型文件夹复制到imagepy/plugins文件夹下，再次启动ImagePy后即可在菜单栏看到该算法。
If the environment is not configured for the model, just enter its menus folder, and run:
pip install -r requirements.txt
PaddleOCR aims to create multilingual, awesome, leading, and practical OCR tools that help users train better models and apply them into practice.
YOLOv5 is a family of compound-scaled object detection models trained on the COCO dataset.
RVM is specifically designed for robust human video matting. Unlike existing neural models that process frames as independent images, RVM uses a recurrent neural network to process videos with temporal memory. RVM can perform matting in real-time on any videos without additional inputs. It achieves 4K 76FPS and HD 104FPS on an Nvidia GTX 1080 Ti GPU.
FBA Matting is an interactive matting. It needs a trimap as an input, which can be easily realized with the ImagePy software.
U2-Net does not need a trimap to detect the salient object in the image.
InsightFace is an open source 2D&3D deep face analysis toolbox, and efficiently implements a rich variety of state of the art algorithms of face recognition, face detection and face alignment, which optimized for both training and deployment.
SimSwap is An Efficient Framework For High Fidelity Face Swapping.
U2-Net can also be used for human portrait drawing.
Cellpose is a generalist algorithm for cell and nucleus segmentation.
Real-ESRGAN aims at developing Practical Algorithms for General Image Restoration.
Old Photo Restoration is for Old Photo Restoration via Deep Latent Space Translation. It can be used to restore fuzzy images and even remove scratches.
LaMa is Resolution-robust Large Mask Inpainting with Fourier Convolutions.
DeepLab is a state-of-art deep learning model for semantic image segmentation, where the goal is to assign semantic labels (e.g., person, dog, cat and so on) to every pixel in the input image.
