Official code for "Towards Photo-Realistic Virtual Try-On by Adaptively Generating↔Preserving Image Content"，CVPR‘20 https://arxiv.org/abs/2003.05863

A clothing retrieval and visual recommendation model for fashion images.

[ECCV 2020] XingGAN for Person Image Generation

[BMVC 2020 Oral] Bipartite Graph Reasoning GANs for Person Image Generation

[CVPR 2022] Exploring Dual-task Correlation for Pose Guided Person Image Generation

🔥 [PR 2023] Multi-scale Attention Guided Pose Transfer (official code).

Fashion editing on Pytorch (python)

pytorch implementation of the deepfashion architecture (https://www.cv-foundation.org/openaccess/content_cvpr_2016/papers/Liu_DeepFashion_Powering_Robust_CVPR_2016_paper.pdf)

🔥 [ECCV 2022] TIPS: Text-Induced Pose Synthesis (official code).

Image retrieval with triplet loss

[Failed] AI to recognize emotions, apparel, body pose and activity in images, and measure how they affect audience engagement. ie. likes and comments.

A deep learning model made on top of the ResNet-50 model for predicting the category of a clothing piece from an image.

DeepFashion Image Classification is a project for deep learning-based clothing recognition using the DeepFashion dataset. It showcases reproducible pipelines and modular models (CNNs, ResNet, ViT, and more) for robust category classification. Designed for learning and experimentation within the fashion tech space.

Projeto final da disciplina EEL7513/EEL7514/EEL510417 - Introdução ao Aprendizado de Máquina.

Experiment with Centroid Re-ID by Changing Backbone Using EfficientNet-V2 and Adding Re-Identification Based Data Augmentation

Generative Adversarial Network on DeepFashion Dataset at a base level

A companion repo to my work done in Cybint

Flask Web App for ES-654 Machine Learning course project

This project builds a hybrid and visual recommender system for (luxury) fashion products, simulating a real-world e-commerce scenario (e.g., Zalando, Farfetch, Fashion Nova). It integrates visual filtering, collaborative filtering, content-based filtering, and hybrid models, enriched with user metadata and interaction recency.