aniketmaster

Detect handwritten words (neural network based).

Official implementation for ICDAR 2021 paper "Handwritten Mathematical Expression Recognition with Bidirectionally Trained Transformer"

꼼꼼한 딥러닝 논문 리뷰와 코드 실습

Document Text Recognition (DocTR) made seamless, high-performing & accessible to anyone using Deep Learning for OCR-related tasks.

[kaggle] 3rd place solution 🌿

ISTR: End-to-End Instance Segmentation with Transformers (https://arxiv.org/abs/2105.00637)

Minimal keyword extraction with BERT

[WACV2021] Oriented Object Detection in Aerial Images with Box Boundary-Aware Vectors

[CVIU 2019] ASSD learns to highlight useful regions on the feature maps while suppressing the irrelevant information, thereby providing reliable guidance for object detection.

More readable and flexible yolov5 with more backbone(resnet, shufflenet, moblienet, efficientnet, hrnet) and (cbam)

Reformer, the efficient Transformer, in Pytorch

YOLOv5 in PyTorch > ONNX > CoreML > TFLite

Self-supervised vIsion Transformer (SiT)

Unofficial Repository for the code of "Detection Transformer" performing object detection on Multi-Mnist dataset.

Evaluation of the Layoutlm model on the CORD dataset