PGNet(Pyramid Graph Networks)[1] modeled structured segmentation data with graphs and further proposed a pyramid-like structure that models different sizes of image regions as graph nodes. There are some details of reading and implementing it.

CRNet(Cross-Reference Networks)[1] make predictions for both the support image and the query image. It can better find the co-occurrent objects in the two images, thus helping the few-shot segmentation task. There are some details of reading and implementing it.

FSL-Survey[1] is a survey on Few-Shot Learning(FSL), which cotains 166 paper to review Few-Shot Learning. They categorize FSL methods from three perspectives: data, model and algorithm. There are some details of reading it.

PV-RCNN[1] is a 3D Object Detection framework to integrate 3D voxel CNN and PointNet-based set abstraction to learn more discriminative point cloud features. The most contributions in this papar is two-stage strategy including the voxel-to-keypoint 3D scene encoding and the keypoint-to-grid RoI feature abstraction. There are some details of reading and implementing it.

DeepSORT[1] integrates appearance information to improve the performance of SORT, learned a deep association metric. There are some details of reading and implementing it.

SORT[1] is pragmatic approach for online and realtime applications. It achieves SOTA with using Kalman filter and Hungarian algorithm. There are some details of reading and implementing it.

FairMOT[1] is a one-shot tracker to fuse object detection and re-identification in a single network. The most contributions in this papar are anchor-free Re-ID feture extraction, multi-layer feature aggregation and lower-dimensional re-ID fetures. There are some details of reading and implementing it.

TSDM[1] is a RGB-D tracker which use depth information to pretreatment and fuse information to pro-processing. It is composed of a Mask-generator(M-g), SiamRPN++ and a Depth-refiner(D-r). There are some details of reading and implementing it.

Image Transformer[1] is a sequence modeling formulation of image generation generalized by Transformer, which restricting the self-attention mechanism to attend to local neighborhoods, while maintaining large receptive field. There are some details of reading and implementing it.

SiamRPN++[1] is a novel Siamese network based tracker to adopt deep networks that broke strict translation invariance. It performs layer-wise and depth-wise aggregations to successfully trained a ResNet-driven Siamese tracker. There are some details of reading and implementing it.