ee67799032e2974486584492039327eb appsci_a2025m1v15n1p350.pdf 77be6fc7fbd4239ca4d84ec318ad00943d646252 appsci_a2025m1v15n1p350.pdf db140ae4bb77eff866d1bb499fe1561c6b78d55f9d821a81ba60aa7a031de3c3 appsci_a2025m1v15n1p350.pdf Title: Exploring Gait Recognition in Wild Nighttime Scenes Subject: Currently, gait recognition research is gradually expanding from ideal indoor environments to real-world outdoor scenarios. However, recognition scenarios in practical applications are often more complex than those considered in existing studies. For instance, real-world scenarios present multiple influencing factors, such as viewpoint variations and diverse carried items. Notably, many gait recognition tasks occur under low-light conditions at night. At present, research on gait recognition in nocturnal environments is relatively limited, and effective methods for nighttime gait recognition are lacking. To address this gap, this study extends gait recognition research to outdoor nighttime environments and introduces the first wild gait dataset encompassing both daytime and nighttime data, named Gait Recognition of Day and Night (GaitDN). Furthermore, to tackle the challenges posed by low-light conditions and other influencing factors in outdoor nighttime gait recognition, we propose a novel pose-based gait recognition framework called GaitSAT. This framework models the intrinsic correlations of human joints by integrating self-attention and graph convolution modules. We conduct a comprehensive evaluation of the proposed method and existing approaches using both the GaitDN dataset and other available datasets. The proposed GaitSAT achieves state-of-the-art performance on the OUMVLP, GREW, Gait3D, and GaitDN datasets, with Rank-1 accuracies of 60.77% , 57.37%, 22.90%, and 86.24%, respectively. Experimental results demonstrate that GaitSAT achieves higher accuracy and superior generalization capabilities compared to state-of-the-art pose-based methods. Keywords: gait recognition; gait recognition dataset; graph convolution networks; temporal; convolutional network; deep learning Author: Haotian Li, Wenjuan Gong, Yutong Li, Yikai Wu, Kechen Li and Jordi Gonzàlez Creator: LaTeX with hyperref Producer: pdfTeX-1.40.25 CreationDate: Thu Jan 2 05:04:25 2025 CET ModDate: Thu Jan 2 05:07:51 2025 CET Custom Metadata: no Metadata Stream: no Tagged: no UserProperties: no Suspects: no Form: none JavaScript: no Pages: 21 Encrypted: no Page size: 595.276 x 841.89 pts (A4) Page rot: 0 File size: 4540832 bytes Optimized: no PDF version: 1.7 name type encoding emb sub uni object ID ------------------------------------ ----------------- ---------------- --- --- --- --------- IJTUFL+URWPalladioL-Roma Type 1 Custom yes yes yes 10 0 TAFCBV+URWPalladioL-Bold Type 1 Custom yes yes yes 16 0 SFOXTA+URWPalladioL-Ital Type 1 Custom yes yes yes 21 0 CRVMKC+TimesNewRomanPS-BoldItalicMT Type 1C WinAnsi yes yes no 33 0 GIGFZE+CMR10 Type 1 Builtin yes yes yes 68 0 PMKYZK+CMSY10 Type 1 Builtin yes yes yes 73 0 QDTWCG+MSBM10 Type 1 Builtin yes yes yes 78 0 QFLWGV+CMEX10 Type 1 Builtin yes yes yes 87 0 SUZJEW+PazoMath-Italic Type 1 Builtin yes yes yes 92 0 QQXUPP+PazoMath Type 1 Builtin yes yes yes 97 0 RXNEKG+MSAM10 Type 1 Builtin yes yes yes 113 0 OLMIHR+SFRM1000 Type 1 Custom yes yes yes 135 0 Jhove (Rel. 1.28.0, 2023-05-18) Date: 2026-03-26 04:11:49 CET RepresentationInformation: appsci_a2025m1v15n1p350.pdf ReportingModule: PDF-hul, Rel. 1.12.4 (2023-03-16) LastModified: 2026-03-25 14:55:24 CET Size: 4540832 Format: PDF Version: 1.7 Status: Well-Formed and valid SignatureMatches: PDF-hul MIMEtype: application/pdf PDFMetadata: Objects: 304 FreeObjects: 1 IncrementalUpdates: 0 DocumentCatalog: PageLayout: SinglePage PageMode: UseNone Outlines: Item: Title: Introduction Destination: section.1 Item: Title: Related Works Destination: section.2 Children: Item: Title: Gait Recognition Datasets Destination: subsection.2.1 Item: Title: Gait Recognition Methods Destination: subsection.2.2 Item: Title: GaitSAT Framework Destination: section.3 Children: Item: Title: Spatial Feature Extraction Destination: subsection.3.1 Item: Title: Multi-Scale Temporal Rhythm Perception Module Destination: subsection.3.2 Item: Title: Network Structure Destination: subsection.3.3 Item: Title: The GaitDN Dataset Destination: section.4 Item: Title: Results and Discussion Destination: section.5 Children: Item: Title: Experimental Settings Destination: subsection.5.1 Item: Title: Ablation Study Destination: subsection.5.2 Children: Item: Title: Analysis of Input Feature Types Destination: subsubsection.5.2.1 Item: Title: Analysis of the Spatial Feature Extraction Module Destination: subsubsection.5.2.2 Item: Title: Analysis of Multi-Scale Temporal Rhythm Perception Module Destination: subsubsection.5.2.3 Item: Title: Performance Comparison with State-of-the-Art Methods Destination: subsection.5.3 Item: Title: Analysis of Generalization Performance Destination: subsection.5.4 Item: Title: Conclusions Destination: section.6 Item: Title: References Destination: section.7 Info: Title: Exploring Gait Recognition in Wild Nighttime Scenes Author: Haotian Li, Wenjuan Gong, Yutong Li, Yikai Wu, Kechen Li and Jordi Gonzàlez Subject: Currently, gait recognition research is gradually expanding from ideal indoor environments to real-world outdoor scenarios. However, recognition scenarios in practical applications are often more complex than those considered in existing studies. For instance, real-world scenarios present multiple influencing factors, such as viewpoint variations and diverse carried items. Notably, many gait recognition tasks occur under low-light conditions at night. At present, research on gait recognition in nocturnal environments is relatively limited, and effective methods for nighttime gait recognition are lacking. To address this gap, this study extends gait recognition research to outdoor nighttime environments and introduces the first wild gait dataset encompassing both daytime and nighttime data, named Gait Recognition of Day and Night (GaitDN). Furthermore, to tackle the challenges posed by low-light conditions and other influencing factors in outdoor nighttime gait recognition, we propose a novel pose-based gait recognition framework called GaitSAT. This framework models the intrinsic correlations of human joints by integrating self-attention and graph convolution modules. We conduct a comprehensive evaluation of the proposed method and existing approaches using both the GaitDN dataset and other available datasets. The proposed GaitSAT achieves state-of-the-art performance on the OUMVLP, GREW, Gait3D, and GaitDN datasets, with Rank-1 accuracies of 60.77% , 57.37%, 22.90%, and 86.24%, respectively. Experimental results demonstrate that GaitSAT achieves higher accuracy and superior generalization capabilities compared to state-of-the-art pose-based methods. Keywords: gait recognition; gait recognition dataset; graph convolution networks; temporal; convolutional network; deep learning Creator: LaTeX with hyperref Producer: pdfTeX-1.40.25 CreationDate: Thu Jan 02 05:04:25 CET 2025 ModDate: Thu Jan 02 05:07:51 CET 2025 ID: 0x72572cfab52774eb90a94ca6a98abc31, 0x72572cfab52774eb90a94ca6a98abc31 Filters: FilterPipeline: FlateDecode Images: Image: NisoImageMetadata: FormatName: image/png ImageWidth: 1476 ImageHeight: 559 ColorSpace: RGB BitsPerSample: 8 BitsPerSampleUnit: integer Filter: FlateDecode Image: NisoImageMetadata: FormatName: image/png ImageWidth: 1256 ImageHeight: 1008 ColorSpace: RGB BitsPerSample: 8 BitsPerSampleUnit: integer Filter: FlateDecode Image: NisoImageMetadata: FormatName: image/png ImageWidth: 894 ImageHeight: 712 ColorSpace: RGB BitsPerSample: 8 BitsPerSampleUnit: integer Filter: FlateDecode Image: NisoImageMetadata: FormatName: image/png ImageWidth: 1390 ImageHeight: 822 ColorSpace: RGB BitsPerSample: 8 BitsPerSampleUnit: integer Filter: FlateDecode Image: NisoImageMetadata: FormatName: image/png ImageWidth: 1569 ImageHeight: 524 ColorSpace: RGB BitsPerSample: 8 BitsPerSampleUnit: integer Filter: FlateDecode Image: NisoImageMetadata: FormatName: image/png ImageWidth: 1368 ImageHeight: 638 ColorSpace: RGB BitsPerSample: 8 BitsPerSampleUnit: integer Filter: FlateDecode Image: NisoImageMetadata: FormatName: image/png ImageWidth: 1237 ImageHeight: 675 ColorSpace: RGB BitsPerSample: 8 BitsPerSampleUnit: integer Filter: FlateDecode Fonts: Type1: Font: BaseFont: TAFCBV+URWPalladioL-Bold FontSubset: true FirstChar: 3 LastChar: 224 FontDescriptor: FontName: TAFCBV+URWPalladioL-Bold Flags: Symbolic FontBBox: -152, -301, 1000, 935 FontFile: true EncodingDictionary: Differences: true ToUnicode: true Font: BaseFont: QQXUPP+PazoMath FontSubset: true FirstChar: 229 LastChar: 229 FontDescriptor: FontName: QQXUPP+PazoMath Flags: Symbolic FontBBox: -40, -283, 878, 946 FontFile: true ToUnicode: true Font: BaseFont: RXNEKG+MSAM10 FontSubset: true FirstChar: 88 LastChar: 88 FontDescriptor: FontName: RXNEKG+MSAM10 Flags: Symbolic FontBBox: 8, -463, 1331, 1003 FontFile: true ToUnicode: true Font: BaseFont: GIGFZE+CMR10 FontSubset: true FirstChar: 40 LastChar: 61 FontDescriptor: FontName: GIGFZE+CMR10 Flags: Symbolic FontBBox: -40, -250, 1009, 750 FontFile: true ToUnicode: true Font: BaseFont: SFOXTA+URWPalladioL-Ital FontSubset: true FirstChar: 45 LastChar: 122 FontDescriptor: FontName: SFOXTA+URWPalladioL-Ital Flags: Symbolic FontBBox: -170, -305, 1010, 941 FontFile: true EncodingDictionary: Differences: true ToUnicode: true Font: BaseFont: QFLWGV+CMEX10 FontSubset: true FirstChar: 0 LastChar: 17 FontDescriptor: FontName: QFLWGV+CMEX10 Flags: Symbolic FontBBox: -24, -2960, 1454, 772 FontFile: true ToUnicode: true Font: BaseFont: OLMIHR+SFRM1000 FontSubset: true FirstChar: 25 LastChar: 25 FontDescriptor: FontName: OLMIHR+SFRM1000 Flags: Symbolic FontBBox: -189, -321, 1456, 937 FontFile: true EncodingDictionary: Differences: true ToUnicode: true Font: BaseFont: PMKYZK+CMSY10 FontSubset: true FirstChar: 0 LastChar: 112 FontDescriptor: FontName: PMKYZK+CMSY10 Flags: Symbolic FontBBox: -29, -960, 1116, 775 FontFile: true ToUnicode: true Font: BaseFont: IJTUFL+URWPalladioL-Roma FontSubset: true FirstChar: 2 LastChar: 246 FontDescriptor: FontName: IJTUFL+URWPalladioL-Roma Flags: Symbolic FontBBox: -166, -283, 1021, 943 FontFile: true EncodingDictionary: Differences: true ToUnicode: true Font: BaseFont: SUZJEW+PazoMath-Italic FontSubset: true FirstChar: 115 LastChar: 115 FontDescriptor: FontName: SUZJEW+PazoMath-Italic Flags: Symbolic FontBBox: -70, -277, 902, 733 FontFile: true ToUnicode: true Font: BaseFont: QDTWCG+MSBM10 FontSubset: true FirstChar: 82 LastChar: 82 FontDescriptor: FontName: QDTWCG+MSBM10 Flags: Symbolic FontBBox: -55, -420, 2343, 920 FontFile: true ToUnicode: true Pages: Page: Sequence: 1 Annotations: Annotation: Subtype: Link Rect: 107, 177, 152, 186 Annotation: Subtype: Link Rect: 34, 165, 129, 174 Annotation: Subtype: Link Rect: 36, 75, 152, 84 Annotation: Subtype: Link Rect: 34, 63, 89, 72 Annotation: Subtype: Link Rect: 34, 773, 175, 809 Annotation: Subtype: Link Rect: 515, 773, 560, 809 Annotation: Subtype: Link Rect: 89, 684, 100, 698 Annotation: Subtype: Link Rect: 185, 684, 196, 698 Annotation: Subtype: Link Rect: 250, 684, 262, 698 Annotation: Subtype: Link Rect: 468, 684, 479, 698 Annotation: Subtype: Link Rect: 422, 27, 560, 37 Annotation: Subtype: Link Rect: 34, 296, 87, 315 Page: Sequence: 2 Annotations: Annotation: Subtype: Link Rect: 514, 703, 521, 712 ActionDest: 147 Annotation: Subtype: Link Rect: 281, 661, 288, 670 ActionDest: 147 Annotation: Subtype: Link Rect: 206, 644, 213, 656 ActionDest: 47 Annotation: Subtype: Link Rect: 518, 619, 525, 628 ActionDest: 147 Annotation: Subtype: Link Rect: 191, 388, 198, 397 ActionDest: 147 Annotation: Subtype: Link Rect: 201, 388, 207, 397 ActionDest: 147 Annotation: Subtype: Link Rect: 196, 301, 204, 313 ActionDest: 47 Annotation: Subtype: Link Rect: 326, 120, 333, 132 ActionDest: 51 Annotation: Subtype: Link Rect: 480, 92, 487, 105 ActionDest: 55 Page: Sequence: 3 Annotations: Annotation: Subtype: Link Rect: 517, 758, 524, 767 ActionDest: 147 Annotation: Subtype: Link Rect: 525, 759, 532, 767 ActionDest: 147 Annotation: Subtype: Link Rect: 483, 120, 490, 130 ActionDest: 147 Annotation: Subtype: Link Rect: 197, 93, 204, 102 ActionDest: 147 Annotation: Subtype: Link Rect: 205, 93, 217, 102 ActionDest: 147 Page: Sequence: 4 Annotations: Annotation: Subtype: Link Rect: 223, 744, 230, 754 ActionDest: 147 Annotation: Subtype: Link Rect: 231, 745, 243, 754 ActionDest: 147 Page: Sequence: 5 Annotations: Annotation: Subtype: Link Rect: 332, 616, 339, 625 ActionDest: 147 Annotation: Subtype: Link Rect: 340, 616, 352, 625 ActionDest: 147 Annotation: Subtype: Link Rect: 443, 616, 450, 625 ActionDest: 147 Annotation: Subtype: Link Rect: 451, 616, 463, 625 ActionDest: 147 Annotation: Subtype: Link Rect: 463, 616, 475, 625 ActionDest: 147 Annotation: Subtype: Link Rect: 414, 574, 426, 583 ActionDest: 147 Annotation: Subtype: Link Rect: 331, 560, 337, 569 ActionDest: 147 Annotation: Subtype: Link Rect: 340, 393, 347, 402 ActionDest: 147 Annotation: Subtype: Link Rect: 220, 309, 232, 318 ActionDest: 147 Annotation: Subtype: Link Rect: 232, 309, 244, 318 ActionDest: 147 Annotation: Subtype: Link Rect: 510, 296, 522, 304 ActionDest: 147 Annotation: Subtype: Link Rect: 367, 102, 379, 111 ActionDest: 147 Annotation: Subtype: Link Rect: 409, 88, 416, 97 ActionDest: 147 Annotation: Subtype: Link Rect: 277, 74, 289, 83 ActionDest: 150 Annotation: Subtype: Link Rect: 410, 74, 422, 83 ActionDest: 150 Page: Sequence: 6 Annotations: Annotation: Subtype: Link Rect: 537, 744, 544, 754 ActionDest: 147 Annotation: Subtype: Link Rect: 544, 744, 556, 754 ActionDest: 147 Annotation: Subtype: Link Rect: 444, 717, 451, 726 ActionDest: 147 Annotation: Subtype: Link Rect: 451, 717, 463, 726 ActionDest: 150 Annotation: Subtype: Link Rect: 466, 717, 478, 726 ActionDest: 150 Annotation: Subtype: Link Rect: 545, 689, 556, 698 ActionDest: 150 Annotation: Subtype: Link Rect: 347, 661, 359, 670 ActionDest: 150 Annotation: Subtype: Link Rect: 519, 619, 530, 628 ActionDest: 150 Annotation: Subtype: Link Rect: 262, 577, 274, 586 ActionDest: 150 Annotation: Subtype: Link Rect: 381, 549, 388, 559 ActionDest: 147 Annotation: Subtype: Link Rect: 307, 508, 319, 517 ActionDest: 150 Annotation: Subtype: Link Rect: 371, 382, 383, 391 ActionDest: 150 Annotation: Subtype: Link Rect: 351, 368, 363, 377 ActionDest: 150 Annotation: Subtype: Link Rect: 426, 327, 438, 336 ActionDest: 150 Annotation: Subtype: Link Rect: 502, 327, 514, 336 ActionDest: 150 Annotation: Subtype: Link Rect: 243, 299, 255, 308 ActionDest: 150 Annotation: Subtype: Link Rect: 245, 229, 257, 238 ActionDest: 150 Page: Sequence: 7 Annotations: Annotation: Subtype: Link Rect: 291, 478, 303, 487 ActionDest: 150 Annotation: Subtype: Link Rect: 547, 476, 554, 487 ActionDest: 65 Page: Sequence: 8 Annotations: Annotation: Subtype: Link Rect: 196, 420, 203, 432 ActionDest: 102 Annotation: Subtype: Link Rect: 428, 409, 435, 418 ActionDest: 147 Annotation: Subtype: Link Rect: 435, 409, 447, 418 ActionDest: 147 Annotation: Subtype: Link Rect: 254, 323, 261, 335 ActionDest: 102 Annotation: Subtype: Link Rect: 369, 116, 381, 126 ActionDest: 150 Annotation: Subtype: Link Rect: 486, 72, 493, 84 ActionDest: 65 Page: Sequence: 9 Page: Sequence: 10 Annotations: Annotation: Subtype: Link Rect: 376, 714, 383, 726 ActionDest: 102 Annotation: Subtype: Link Rect: 362, 160, 369, 169 ActionDest: 110 Page: Sequence: 11 Annotations: Annotation: Subtype: Link Rect: 496, 512, 508, 521 ActionDest: 150 Annotation: Subtype: Link Rect: 332, 456, 344, 465 ActionDest: 150 Annotation: Subtype: Link Rect: 335, 400, 347, 410 ActionDest: 150 Annotation: Subtype: Link Rect: 519, 273, 526, 285 ActionDest: 51 Annotation: Subtype: Link Rect: 440, 259, 446, 271 ActionDest: 119 Page: Sequence: 12 Annotations: Annotation: Subtype: Link Rect: 407, 184, 414, 193 ActionDest: 123 Page: Sequence: 13 Annotations: Annotation: Subtype: Link Rect: 532, 714, 539, 726 ActionDest: 123 Annotation: Subtype: Link Rect: 348, 489, 355, 501 ActionDest: 123 Page: Sequence: 14 Annotations: Annotation: Subtype: Link Rect: 189, 714, 201, 723 ActionDest: 150 Annotation: Subtype: Link Rect: 201, 714, 213, 723 ActionDest: 150 Annotation: Subtype: Link Rect: 214, 714, 226, 723 ActionDest: 150 Annotation: Subtype: Link Rect: 226, 714, 238, 723 ActionDest: 150 Annotation: Subtype: Link Rect: 359, 697, 366, 709 ActionDest: 126 Annotation: Subtype: Link Rect: 464, 644, 476, 653 ActionDest: 150 Annotation: Subtype: Link Rect: 230, 526, 242, 535 ActionDest: 150 Annotation: Subtype: Link Rect: 229, 509, 241, 518 ActionDest: 150 Annotation: Subtype: Link Rect: 234, 492, 246, 501 ActionDest: 150 Annotation: Subtype: Link Rect: 228, 475, 240, 484 ActionDest: 150 Annotation: Subtype: Link Rect: 345, 367, 352, 379 ActionDest: 129 Annotation: Subtype: Link Rect: 277, 175, 284, 184 ActionDest: 147 Annotation: Subtype: Link Rect: 285, 175, 296, 184 ActionDest: 147 Page: Sequence: 15 Annotations: Annotation: Subtype: Link Rect: 256, 663, 264, 673 ActionDest: 147 Annotation: Subtype: Link Rect: 304, 593, 318, 602 ActionDest: 147 Annotation: Subtype: Link Rect: 508, 318, 520, 328 ActionDest: 150 Annotation: Subtype: Link Rect: 301, 188, 308, 200 ActionDest: 132 Page: Sequence: 16 Annotations: Annotation: Subtype: Link Rect: 443, 686, 450, 698 ActionDest: 132 Annotation: Subtype: Link Rect: 413, 136, 420, 148 ActionDest: 141 Page: Sequence: 17 Page: Sequence: 18 Annotations: Annotation: Subtype: Link Rect: 470, 661, 482, 670 ActionDest: 150 Annotation: Subtype: Link Rect: 485, 661, 497, 670 ActionDest: 153 Annotation: Subtype: Link Rect: 261, 285, 519, 296 Annotation: Subtype: Link Rect: 343, 272, 560, 283 Annotation: Subtype: Link Rect: 165, 259, 235, 270 Annotation: Subtype: Link Rect: 444, 259, 544, 270 Annotation: Subtype: Link Rect: 366, 246, 560, 257 Annotation: Subtype: Link Rect: 165, 233, 212, 244 Page: Sequence: 19 Annotations: Annotation: Subtype: Link Rect: 393, 611, 431, 622 Annotation: Subtype: Link Rect: 59, 471, 96, 481 Annotation: Subtype: Link Rect: 102, 471, 138, 481 Annotation: Subtype: Link Rect: 305, 343, 342, 353 Annotation: Subtype: Link Rect: 349, 343, 385, 353 Annotation: Subtype: Link Rect: 434, 253, 472, 264 Annotation: Subtype: Link Rect: 484, 228, 521, 239 Page: Sequence: 20 Annotations: Annotation: Subtype: Link Rect: 267, 744, 305, 754 Annotation: Subtype: Link Rect: 311, 744, 347, 754 Annotation: Subtype: Link Rect: 202, 719, 239, 729 Annotation: Subtype: Link Rect: 246, 719, 282, 729 Annotation: Subtype: Link Rect: 163, 539, 201, 550 Annotation: Subtype: Link Rect: 233, 450, 270, 461 Annotation: Subtype: Link Rect: 91, 144, 128, 154 Annotation: Subtype: Link Rect: 134, 144, 170, 154 Annotation: Subtype: Link Rect: 322, 92, 360, 103 Page: Sequence: 21 Annotations: Annotation: Subtype: Link Rect: 368, 743, 406, 754 Annotation: Subtype: Link Rect: 118, 719, 155, 729 Checksum: fc070363 Type: CRC32 Checksum: ee67799032e2974486584492039327eb Type: MD5 Checksum: 77be6fc7fbd4239ca4d84ec318ad00943d646252 Type: SHA-1 Checksum: db140ae4bb77eff866d1bb499fe1561c6b78d55f9d821a81ba60aa7a031de3c3 Type: SHA-256