Tag：paper
-
Time：2021-10-3
1. Program structure The experiment of lab2 is to realize the following interfaces // create a new Raft server instance: rf := Make(peers, me, persister, applyCh) // start agreement on a new log entry: rf.Start(command interface{}) (index, term, isleader) // ask a Raft for its current term, and whether it thinks it is leader rf.GetState() […]
-
Time：2021-9-30
CIKM is a top international academic conference in the fields of information retrieval, knowledge management and database. Since 1992, CIKM has successfully brought together first-class researchers and developers in the above three fields, providing an international forum for exchanging the latest developments in information and knowledge management research, data and knowledge base. The purpose of […]
-
Time：2021-9-17
This paper proposes that Pconv can effectively integrate the internal relationship between scales by 3D convolution of feature pyramid and regularization with specific Ibn. In addition, this paper proposes SEPC, which uses deformable convolution to adapt to the irregularity of the correspondence between actual features and maintain scale equilibrium. Pconv and SEPC significantly improve the […]
-
Time：2021-8-27
Content summary: a research on the automatic translation of cartoon characters has aroused heated discussion. The mantra team composed of two doctors from the University of Tokyo has released a paper, which has been included by AAAI 2021. The mantra project aims to provide automatic machine translation tools for Japanese comics. Original: hyperai hypernerve Key […]
-
Time：2021-8-12
From the perspective of theory, this paper makes an in-depth study on the domain adaptation of target detection. Based on the countermeasure training of h-diversity, Da fast r-cnn is proposed, which aligns the domain from the perspective of picture level and instance level, and adds consistent regularization to learn the domain invariant RPN. From the […]
-
Time：2021-8-9
Based on hierarchical expression, this paper proposes an efficient evolutionary algorithm to search the neural network structure, and constructs a powerful convolution structure by stacking layers. The search method of this paper is simple. From the experimental results, it achieves a very good accuracy and is worth learning Source: Xiao Fei’s algorithm Engineering Notes] official […]
-
Time：2021-8-7
This paper is published by the author of centernet. This paper proposes anchor free / two-stage target detection algorithm CPN, which uses key points to extract candidate boxes, and then uses two-stage classifier for prediction. The overall idea of the paper is very simple, but the accuracy and reasoning speed of CPN are very good. […]
-
Time：2021-8-4
Based on the key point prediction network, this paper proposes centernet algorithm, which regards the detected target as the key point, finds the center point of the target first, and then returns its size. Compared with the previous centernet algorithm with the same name, the algorithm in this paper is more concise and powerful enough. […]
-
Time：2021-7-31
This paper proposes aging evolution, a variant of tournament selection to optimize the evolutionary algorithm. In the nasnet search space, compared with reinforcement learning and random search, the algorithm is simple enough and can search higher quality models faster. The amoebanet-a searched in this paper can reach SOTA on Imagenet Source: Xiao Fei’s algorithm Engineering […]
-
Time：2021-7-30
As an anchor free paper in the same period as fcos and fsaf, foveabox is also based on the densebox plus FPN strategy in the overall structure. The main difference is that foveabox only uses the target central area for prediction, and the regression prediction is the normalized offset value. In addition, multiple layers of […]
-
Time：2021-7-30
Squeezenet series is an early and classic lightweight network. Squeezenet uses fire module for parameter compression, and squeezenext adds separation convolution to improve it. Although the squeezenet series is not as widely used as mobinet, its architecture idea and experimental conclusions can be used for reference. Source: Xiaofei’s algorithm Engineering Notes official account SqueezeNet Paper: […]
-
Time：2021-7-29
Guided anchoring solves the problems of conventional manual preset anchors by generating anchors online, and provides two implementation methods for embedding according to the adaptive characteristics of the generated anchors. It is a very complete solution Source: Xiaofei’s algorithm Engineering Notes official account Thesis: region proposal by guided anchorage Thesis address:https://arxiv.org/abs/1901.03278 Thesis Code:https://github.com/open-mmlab/mmdetection Introduction Anchor […]
