Tag：paper
-
Time：2021-7-27
Thesis Name: the evolution of the UNIX time sharing system — Danis M. Ritchie Dennis rich reviewed the process of creating UNIX with Ken Thompson at Bell Labs, and then introduced the most important parts of UNIX, including: file system I-list: an array of i-nodes. Each i-node describes a file. I-node contains the metadata of […]
-
Time：2021-7-27
Compile | CVReport | opencv Chinese website (wechat ID: iopencv) This article reviews the papers related to wacv2021 image segmentation, including matting, instance, panorama, semantic segmentation, natural disaster assessment and other related applications. It is noteworthy that there is a text matting that has been rarely or never studied in previous work. A total of […]
-
Time：2021-7-24
Shufflenet series is a very important series in lightweight networks. Shufflenetv1 puts forward channel shuffle operation, so that the network can use packet convolution to accelerate, while shufflenetv2 pushes down most of the designs of V1, puts forward channel split operation from reality, accelerates the network and reuses features at the same time, and achieves […]
-
Time：2021-7-23
Peleenet is a variant of densenet, which does not use the popular deep separable convolution. Peleenet and Pelee only achieve good performance and speed through structural optimization. After reading this paper, we can learn a lot of network design tips. Source: Xiaofei’s algorithm Engineering Notes official account Paper: Pelee: a real time object detection system […]
-
Time：2021-7-15
Aiming at the optimization problem of anchor point detection algorithm, this paper proposes SAPd method, which uses different loss weights for anchor points in different locations, and carries out weighted common training for different feature pyramid layers, which eliminates most of the artificial rules and more follows the weights of the network itself for training […]
-
Time：2021-7-8
This paper proposes a neural network architecture search method for mobile terminal. There are two main ideas in this method. Firstly, the multi-objective optimization method is used to integrate the time-consuming of the model on the actual device into the search, and then the hierarchical search space is used to keep the diversity of the […]
-
Time：2021-7-4
CSP defines the target as the center point and size, and directly predicts the center and size of the target through the network, which is much lighter than the traditional RCNN type detection algorithm. The whole idea collided with object as points. It’s really a hero’s opinion Source: Xiaofei’s algorithm Engineering Notes official account Paper: […]
-
Time：2021-6-29
The characteristic of condensenet is that it can learn the grouping convolution and prune in combination with the training process. It can not only prune accurately, but also continue to train and make the network weight smoother, which is a very good work Source: Xiaofei’s algorithm Engineering Notes official account Thesis: neural architecture search with […]
-
Time：2021-6-24
Extremenet detects the four poles of the target, and then combines them in a geometric way to detect the target. The performance of extremenet is equivalent to other traditional detection algorithms. Extreme net detection method is very unique, but it contains more post-processing methods, so there is a lot of room for improvement. If you […]
-
Time：2021-6-23
The core of centripetalnet is a new corner matching method, which can learn an additional centripetal offset value. The corner with small enough offset value is the matching. Compared with the matching method of embedding vector, this method is more robust and has better interpretability. In addition, the cross star deformation convolution proposed in this […]
-
Time：2021-6-13
The core of IGC series network is the extreme application of packet convolution. The conventional convolution is decomposed into multiple packet convolutions, which can reduce a large number of parameters. In addition, the principle of complementarity and sorting operation can ensure the information flow between packets with the least number of parameters. But on the […]
-
Time：2021-6-11
Today, I’d like to introduce five papers on ECCV 2020 oral. ECCV, CVPR and iccv are called the top three conferences in the field of computer vision. A total of 5025 articles were submitted to ECCV 2020, of which 1361 were accepted, with an acceptance rate of 27%. ECCV 2020 accepted papers list address: https://eccv2020.eu/accepted-papers/ […]
