Kaggle Competition

---

iMaterialist (Fashion) 2020 at FGVC7
Fine-grained segmentation task for fashion and apparel

Description

In this competition, we challenge you to develop algorithms that will help with an important step towards automatic product detection – to accurately assign segmentations and attribute labels for fashion images.

Data

• train.csv
• train imageset
• test imageset
• label_description_json
• sample_submission.csv

Approach

• Segmentation - Mask R CNN
• Augmentation - imgaug library

Result, Segmentation

To Do

• LearningRate Warmup : http://openaccess.thecvf.com/content_CVPR_2019/papers/He_Bag_of_Tricks_for_Image_Classification_with_Convolutional_Neural_Networks_CVPR_2019_paper.pdf
• Attributes Multilabel Calssification : https://towardsdatascience.com/multi-label-image-classification-with-neural-network-keras-ddc1ab1afede

Conclustion