Abstract:The ability to correctly classify and retrieve apparel images has a variety of applications important to e-commerce, online advertising and internet search. In this work, we propose a robust framework for fine-grained apparel classification, in-shop and cross-domain retrieval which eliminates the requirement of rich annotations like bounding boxes and human-joints or clothing landmarks, and training of bounding box/ key-landmark detector for the same. Factors such as subtle appearance differences, variations in human poses, different shooting angles, apparel deformations, and self-occlusion add to the challenges in classification and retrieval of apparel items. Cross-domain retrieval is even harder due to the presence of large variation between online shopping images, usually taken in ideal lighting, pose, positive angle and clean background as compared with street photos captured by users in complicated conditions with poor lighting and cluttered scenes. Our framework uses compact bilinear CNN with tensor sketch algorithm to generate embeddings that capture local pairwise feature interactions in a translationally invariant manner. For apparel classification, we pass the feature embeddings through a softmax classifier, while, the in-shop and cross-domain retrieval pipelines use a triplet-loss based optimization approach, such that squared Euclidean distance between embeddings measures the dissimilarity between the images. Unlike previous works that relied on bounding box, key clothing landmarks or human joint detectors to assist the final deep classifier, proposed framework can be trained directly on the provided category labels or generated triplets for triplet loss optimization. Lastly, Experimental results on the DeepFashion fine-grained categorization, and in-shop and consumer-to-shop retrieval datasets provide a comparative analysis with previous work performed in the domain.

Clothing Retrieval Based on Image Bundled Features

Measuring Clothing Image Similarity with Bundled Features

3D Clothing Fitting Research Based on Feature Matching

Fashion Recommendation on Street Images.

Automatic Measurement of Garment Sizes Using Image Recognition

Snap and Find: Deep Discrete Cross-domain Garment Image Retrieval

An Approach for Clothing Recommendation Based on Multiple Image Attributes

Semantic Classification Based Clothing Image Retrieval

Two-Stage Method for Clothing Feature Detection

Fine-grained Apparel Classification and Retrieval without rich annotations

Intelligent Clothing Matching Based on Feature Analysis

Query-free Clothing Retrieval via Implicit Relevance Feedback

Toward Realistic Virtual Try-on Through Landmark Guided Shape Matching

Cloth Interactive Transformer for Virtual Try-On

An End-to-End Framework for Clothing Collocation Based on Semantic Feature Fusion

Clothing Co-Parsing by Joint Image Segmentation and Labeling

Clothing Similarity Computation Based on Tlac

Searching for Apparel Products from Images in the Wild

Learning To Appreciate The Aesthetic Effects Of Clothing

Adding Affine Invariant Geometric Constraint for Partial-Duplicate Image Retrieval

Reconstructing Complex Shaped Clothing from a Single Image with Feature Stable Unsigned Distance Fields