본문 바로가기 주메뉴 바로가기
22대 대한민국 국회 국회정보길라잡이 국회의원검색
회원가입 로그인
HOME

정보검색

소장정보 검색

국회도서관 홈으로 정보검색 소장정보 검색

결과 내 검색

동의어 포함

고급검색

인명/단체명

저자정보 상세정보
인명/단체명을 입력하세요.

키워드

  • 대표어

  • 외국어

-

목차보기

Title Page 2

Contents 5

Abstract 9

Chapter 1. Introduction 10

1.1. Motivation 10

1.2. Overview of the Proposed Methods 13

Chapter 2. Related Works 15

2.1. CNNs 15

2.2. RNNs 18

2.3. GNNs 19

2.4. Transformers 20

Chapter 3. Proposed Methods 22

3.1. State Space Model (SSM) 22

3.2. Proposed Methods 26

Chapter 4. Experiments 33

4.1. Dataset and Evaluation Metrics 33

4.2. Evaluation on Benchmarks 34

4.3. Ablation studies 35

Chapter 5. Conclusion 39

5.1. Conclusion 39

5.2. Limitations and Future work 40

Bibliography 41

List of Tables 8

Table 2.1. Strategies and performance comparison of each fundamental model (CNN, RNN, GCN, Transformer, Mamba) for improving TAL performance on the THUMOS-14... 16

Table 4.1. Results of temporal action localization on benchmark datasets 34

Table 4.2. Ablation studies on the proposed methods 36

List of Figures 7

Figure 3.1. The architecture overview of proposed methods 26

Figure 3.2. Diagrams of the Embedding, Stem, and Branch modules 27

Figure 3.3. Process of extracting spatiotemporal features using the pretrained video encoder 29

Figure 3.4. Diagrams of the Embedding, Stem, and Branch modules 30

초록보기

 Temporal Action Localization (TAL) is crucial for understanding actions within videos by classifying them and determining their temporal boundaries. Traditional deep learning approaches using CNNs, RNNs, GCNs, and Transformers have made significant strides yet often struggle with capturing long-range dependencies (LRD) and efficiently processing extensive video sequences. This thesis derives insights from these methods to enhance TAL performance, specifically by leveraging a combination of Feature aggregated Bi- S6 block design, Dual Bi-S6 structure, and Recurrent mechanism.

Our approach involves three key components. First, the Feature aggregated Bi-S6 block uses multiple Conv1D layers with various kernel sizes in parallel, summing their outputs to capture local contexts of different ranges. This aggregated result is then fed into the Bi-directional S6 (Bi-S6), enhancing its capacity to model complex features. Sec-ond, the Dual Bi-S6 structure employs two parallel Feature Aggregated Bi-S6 blocks: one (TFA-Bi-S6) processes the temporal dimension, and the other (C-Bi-S6) processes the channel dimension. The outputs are then combined using point-wise multiplication, effec-tively integrating temporal dependencies of spatiotemporal features and spatiotemporal dependencies of temporal features, thereby making TAL more robust. Third, the Recur-rent mechanism applies the Dual Bi-S6 structure recursively r times in a residual manner, allowing the model to iteratively refine its representation of long-range dependencies.

The proposed architecture, inspired by ActionFormer and ActionMamba, includes a Pretrained video encoder that extracts spatiotemporal features for each clip of a video. It features a Backbone that captures dependencies and extracts features at various temporal resolutions from the sequence data containing spatiotemporal features of each clip. The architecture also includes a simple post-processing Neck for handling multiple temporal resolutions and a Head that classifies actions and performs regression on video segments using the post-processed results.

Extensive evaluation on benchmark datasets such as THUMOS-14, ActivityNet, Fine- Action, and HACS validates our approach, with our models outperforming existing state-of-

the-art solutions. We achieved remarkable mean Average Precision (mAP) scores of 74.2% on THUMOS-14, 42.9% on ActivityNet, 29.6% on FineAction, and 45.8% on HACS.

추천서가 (다양한 추천 자료를 만나보세요)

권호기사보기

권호기사 목록 테이블로 기사명, 저자명, 페이지, 원문, 기사목차 순으로 되어있습니다.
기사명 저자명 페이지 원문 기사목차

권호

기사명 원문보기 기사목차
닫기