1"""Area perception_interaction, part A: categories 9-11 (computer vision, NLP, speech and audio)."""23CATEGORIES = [4dict(5 id="computer_vision", number=9, name="Computer Vision", area="perception_interaction",6 definition=("Extracting information from images, video and 3D data: recognition, detection, segmentation, tracking, "7 "pose, depth and 3D reconstruction, image restoration and enhancement, medical and remote-sensing image "8 "analysis, visual representation learning. Image synthesis is generative_ai; vision-language is multimodal_ai; "9 "rendering is computer_graphics."),10 topics=[("T10036", 1.0, "object detection, segmentation, YOLO"), ("T10531", 1.0, "advanced vision and imaging"),11 ("T10331", 1.0, "video surveillance and tracking"), ("T11448", 1.0, "face recognition"),12 ("T10057", 1.0, "face and expression recognition"), ("T10812", 1.0, "human pose and action recognition"),13 ("T12549", 1.0, "image and object detection"), ("T11605", 1.0, "visual attention and saliency"),14 ("T14339", 1.0, "image processing and 3D reconstruction"), ("T10601", 0.8, "handwritten text recognition"),15 ("T10824", 0.8, "image retrieval and classification"), ("T11019", 0.8, "image enhancement"),16 ("T11105", 0.8, "advanced image processing"), ("T11439", 0.8, "video analysis and summarization"),17 ("T12923", 0.8, "digital image processing"), ("T13579", 0.8, "image and video stabilization"),18 ("T12111", 0.8, "industrial vision and defect detection"), ("T10052", 0.7, "medical image segmentation"),19 ("T10688", 0.7, "image and signal denoising"), ("T10627", 0.7, "image and video retrieval"),20 ("T11165", 0.7, "image and video quality assessment"), ("T13114", 0.7, "image processing techniques"),21 ("T10444", 0.6, "context-aware activity recognition"), ("T10689", 0.6, "remote-sensing image classification"),22 ("T11659", 0.6, "image fusion"), ("T12422", 0.5, "radiomics and ML in medical imaging"),23 ("T12983", 0.5, "satellite image processing and photogrammetry"), ("T12859", 0.5, "cell image analysis"),24 ("T10719", 0.5, "3D shape modeling and analysis"), ("T11398", 0.5, "hand gesture recognition"),25 ("T14319", 0.5, "currency recognition"), ("T10522", 0.4, "medical imaging techniques"),26 ("T12357", 0.4, "digital media forensics"), ("T11714", 0.4, "multimodal (vision-language)"),27 ("T11707", 0.4, "gaze tracking"), ("T10775", 0.3, "GANs and image synthesis"),28 ("T10888", 0.3, "augmented reality"), ("T10901", 0.3, "advanced data compression (image)"),29 ("T10741", 0.3, "video coding"), ("T10388", 0.3, "steganography and watermarking"),30 ("T11017", 0.2, "chaos-based image encryption"), ("T12874", 0.4, "digital imaging for blood diseases"),31 ("T10638", 0.2, "optical measurement and interference"), ("T13759", 0.3, "medical imaging and pathology")],32 subfields=[(1707, 0.5)],33 arxiv=[("cs.CV", 1.0), ("eess.IV", 0.6)],34 streams=[("conf/cvpr", 1.0), ("conf/iccv", 1.0), ("conf/eccv", 1.0), ("conf/wacv", 1.0), ("conf/bmvc", 1.0),35 ("conf/accv", 1.0), ("conf/iccvw", 1.0), ("conf/fgr", 1.0), ("conf/3dim", 1.0), ("journals/ijcv", 1.0),36 ("journals/cviu", 1.0), ("journals/ivc", 1.0), ("journals/mva", 1.0), ("journals/pami", 0.8),37 ("journals/tip", 0.8), ("conf/icip", 0.8), ("conf/icpr", 0.8), ("conf/icdar", 0.8), ("conf/visapp", 0.8),38 ("conf/prcv", 0.8), ("journals/tcsv", 0.7), ("journals/iet-ipr", 0.7), ("journals/jrtip", 0.7),39 ("journals/pr", 0.6), ("journals/prl", 0.6), ("journals/jvcir", 0.6), ("journals/sivp", 0.6),40 ("journals/jimaging", 0.6), ("conf/miccai", 0.6), ("journals/mia", 0.6), ("journals/tmi", 0.5),41 ("conf/isbi", 0.5), ("journals/paa", 0.5), ("journals/tci", 0.5), ("journals/imst", 0.5),42 ("journals/cmig", 0.5), ("journals/spic", 0.5), ("journals/jei", 0.5), ("journals/cars", 0.4),43 ("journals/bmcmi", 0.4), ("journals/ijprai", 0.5), ("journals/tgrs", 0.3), ("conf/igarss", 0.2),44 ("journals/remotesensing", 0.2), ("journals/lgrs", 0.3), ("journals/staeors", 0.2), ("conf/vcip", 0.6),45 ("conf/icmcs", 0.3), ("conf/mm", 0.2), ("journals/mta", 0.3), ("conf/ivs", 0.3), ("journals/tits", 0.1),46 ("journals/jdi", 0.3), ("journals/jmihi", 0.3), ("conf/bmei", 0.4), ("journals/displays", 0.2),47 ("conf/icpram", 0.5), ("journals/cg", 0.2)],48 keywords=[49 (r"(?i)\bcomputer vision\b|\bmachine vision\b|\bvision (models?|transformers?|tasks?|backbones?|systems?)\b", 1.0, "both"),50 (r"(?i)\b(object|face|facial|pedestrian|vehicle|action|activity|pose|gesture|scene|visual|image|video|person|text|lane|landmark|anomaly in video|salient object) (detection|recognition|tracking|re-?identification|classification|segmentation|understanding|retrieval|estimation|parsing|captioning|matting)\b", 0.8, "both"),51 (r"(?i)\b(semantic|instance|panoptic|medical image|image|video|3D|point[- ]cloud|interactive|few-?shot|open-?vocabulary) segmentation\b", 1.0, "both"),52 (r"(?i)\b(point clouds?|optical flow|depth (estimation|completion)|stereo (matching|vision)|3D reconstruction|structure[- ]from[- ]motion|SfM\b|multi-?view (stereo|geometry)|NeRF|neural radiance fields?|Gaussian splatting|visual (SLAM|odometry|localization)|camera (pose|calibration)|epipolar|homography|feature (matching|descriptors?)|SIFT|keypoints?)\b", 1.0, "both"),53 (r"(?i)\b(image|video) (processing|enhancement|denoising|restoration|deblurring|dehazing|super-?resolution|quality assessment|registration|retrieval|classification|synthesis|inpainting|compression|coding|forensics|fusion|stitching|matting|colori[sz]ation|harmoni[sz]ation)\b|\bvideo (surveillance|understanding|analysis|summari[sz]ation|prediction|anomaly)\b", 0.8, "both"),54 (r"(?i)\b(convolutional neural networks?|CNNs?|YOLO\w*|R-?CNN|Mask R-?CNN|DETR|U-?Net|ViTs?|vision transformers?|Swin|ResNet|VGG|backbone networks?|feature pyramid|anchor-?free|non-?maximum suppression|IoU\b|mAP\b|COCO|ImageNet|Pascal VOC|KITTI|nuScenes|Cityscapes|ADE20K)\b", 0.6, "both"),55 (r"(?i)\b(medical imag(e|es|ing)|radiomics|histopatholog|whole-?slide|retinal (image|fundus)|OCT (image|scan)|MRI|CT (scans?|images?)|X-?ray (images?|classification)|ultrasound (image|video)|mammogra|dermoscop|endoscop|colonoscop|microscopy images?)\b", 0.5, "both"),56 (r"(?i)\b(saliency|visual attention|OCR|optical character recognition|handwrit|scene text|document (image|layout|understanding)|license plate|biometrics?|iris recognition|fingerprint|gait recognition|crowd counting|human parsing|6-?DoF pose|hand pose|body pose|facial (landmarks?|expression|attributes?)|deepfake detection|face anti-?spoofing|remote sensing (images?|scene)|hyperspectral (image|classification)|SAR (images?|target)|change detection in|aerial imag|satellite imag|UAV imag)\b", 0.8, "both"),57 (r"(?i)\b(self-?supervised visual|contrastive visual|visual representation|visual features?|image embeddings?|visual recognition|visual perception|pixel-?wise|pixel-?level|bounding box(es)?|RGB-?D|LiDAR (point|perception|segmentation|detection)|BEV (perception|features?|segmentation)|bird'?s-?eye view|autonomous driving perception|3D (object detection|detection|perception|vision|human))\b", 0.8, "both"),58 (r"(?i)\b(images?|visual|pixels?|cameras?|videos?)\b", 0.2, "both"),59 ],60 exclusions=[],61 related=["multimodal_ai", "computer_graphics", "robotics", "deep_learning_architectures", "generative_ai", "computational_biology"],62 parent_of=[], child_of=[],63 provenance=("All 35 topics of OpenAlex subfield 1707 reviewed; those that are image/video-centred receive 0.5-1.0, "64 "media/compression/encryption topics are shared at low weight; relevant Engineering (22) remote-sensing "65 "topics and Medicine imaging topics added at partial weight. cs.CV is a sufficient signal."),66),67dict(68 id="nlp", number=10, name="Natural Language Processing", area="perception_interaction",69 definition=("Computational processing of human language: machine translation, parsing and tagging, information "70 "extraction, sentiment and text classification, summarisation, question answering, dialogue, text mining, "71 "computational linguistics, low-resource and multilingual NLP, NLP evaluation and resources. LLM-centred "72 "work is llm (also classified here when the task framing dominates)."),73 topics=[("T10181", 1.0, "NLP techniques: MT, parsing, NLU/NLG"), ("T10664", 1.0, "sentiment analysis and opinion mining"),74 ("T11550", 1.0, "text and document classification"), ("T13629", 1.0, "text readability and simplification"),75 ("T13083", 0.8, "advanced text analysis"), ("T12380", 0.8, "authorship attribution and profiling"),76 ("T11710", 0.8, "biomedical text mining and ontologies"), ("T12262", 0.7, "hate speech detection"),77 ("T10028", 0.6, "topic modeling / pre-trained LMs"), ("T12031", 0.6, "speech and dialogue systems"),78 ("T13910", 0.5, "computational and text analysis methods"), ("T13623", 0.4, "linguistic studies and language acquisition"),79 ("T10215", 0.4, "semantic web and ontologies"), ("T13274", 0.4, "expert finding and Q&A"),80 ("T11010", 0.3, "logic, reasoning and knowledge representation"), ("T10034", 0.3, "syntax, semantics, linguistic variation"),81 ("T11644", 0.3, "spam and phishing detection (text)"), ("T11147", 0.3, "misinformation"),82 ("T12016", 0.3, "web data mining and analysis"), ("T13155", 0.3, "digital communication and language"),83 ("T12353", 0.2, "lexicography and language studies"), ("T13310", 0.2, "subtitles and audiovisual media")],84 subfields=[(1702, 0.1)],85 arxiv=[("cs.CL", 1.0), ("cs.IR", 0.1)],86 streams=[("conf/acl", 1.0), ("conf/emnlp", 1.0), ("conf/naacl", 1.0), ("conf/eacl", 1.0), ("conf/coling", 1.0),87 ("journals/tacl", 1.0), ("journals/coling", 1.0), ("conf/lrec", 1.0), ("conf/semeval", 1.0), ("conf/wmt", 1.0),88 ("conf/nlpcc", 1.0), ("journals/talip", 1.0), ("conf/wassa", 1.0), ("conf/sepln", 0.8), ("journals/csl", 0.6),89 ("journals/taslp", 0.5), ("conf/clef", 0.5), ("journals/ipm", 0.3), ("conf/dihu", 0.3), ("conf/dhd", 0.3),90 ("conf/interspeech", 0.2), ("conf/aaai", 0.1), ("conf/ijcai", 0.1), ("journals/kbs", 0.1),91 ("journals/eswa", 0.1), ("conf/cikm", 0.2), ("conf/www", 0.1), ("journals/jasis", 0.2), ("conf/cogsci", 0.2)],92 keywords=[93 (r"(?i)\bnatural language (processing|understanding|generation|inference|interfaces?)\b|\bNLP\b|\bNLU\b|\bNLG\b|\bNLI\b", 1.0, "both"),94 (r"(?i)\b(machine translation|neural machine translation|NMT\b|named entity recognition|NER\b|part-?of-?speech|POS tagging|dependency parsing|constituency parsing|coreference resolution|word embeddings?|word2vec|GloVe|sentiment analysis|opinion mining|aspect-?based sentiment|text (classification|summari[sz]ation|generation|mining|simplification|normali[sz]ation|clustering|retrieval)|abstractive summari[sz]ation|extractive summari[sz]ation|question answering|reading comprehension|relation extraction|event extraction|information extraction|semantic (parsing|role labeling|similarity|textual similarity)|textual entailment|language identification|tokeni[sz]ation|lemmati[sz]ation|stemming|word sense disambiguation|grammatical error correction|spell(ing)? correction|discourse (parsing|analysis)|argument mining|keyphrase extraction|topic model(ing|s)|LDA\b|latent Dirichlet)\b", 1.0, "both"),95 (r"(?i)\b(computational linguistics|corpus linguistics|treebanks?|low-?resource languages?|multilingual|cross-?lingual|code-?switching|morpholog(y|ical) (analysis|segmentation)|lexical (semantics|resources?)|WordNet|language resources?|annotated corp(us|ora)|inter-?annotator)\b", 0.8, "both"),96 (r"(?i)\b(dialogue (systems?|state tracking|generation|management|acts?)|task-?oriented dialogue|open-?domain dialogue|chatbots?|conversational (AI|agents?|systems?|search))\b", 0.7, "both"),97 (r"(?i)\b(hate speech|fake news|misinformation|rumou?r|stance|sarcasm|irony|emotion|humou?r|offensive language|toxic(ity)?|propaganda|clickbait|authorship) (detection|classification|identification|recognition|attribution|analysis)\b", 0.8, "both"),98 (r"(?i)\b(language models?|BERT|GPT|transformers?|sequence-?to-?sequence|seq2seq|encoder-?decoder|attention)\b", 0.3, "both"),99 (r"(?i)\b(knowledge graph|ontolog(y|ies)|semantic web|entity linking|knowledge base) (completion|embedding|construction|population|question answering|alignment|learning)?\b", 0.4, "both"),100 (r"(?i)\b(text|texts|textual|linguistic|lexical|syntactic|semantic|sentences?|words?|documents?|corpus|corpora|language)\b", 0.2, "both"),101 ],102 exclusions=[],103 related=["llm", "speech_audio", "information_retrieval", "multimodal_ai", "computational_biology"],104 parent_of=[], child_of=[],105 provenance=("OpenAlex 1702 language topics (T10181, T10664, T11550, T13083, T12380, T13629, T12262) plus biomedical "106 "text mining (field 13) and social-science text topics. *CL venues are the strongest venue signal in the "107 "whole ontology (ACL Anthology coverage)."),108),109dict(110 id="speech_audio", number=11, name="Speech and Audio Computing", area="perception_interaction",111 definition=("Processing and generation of speech, music and general audio: speech recognition and synthesis, "112 "speaker and language recognition, enhancement and separation, spoken language understanding, music "113 "information retrieval and generation, audio event detection, neural audio codecs. Text-only dialogue is nlp."),114 topics=[("T10201", 1.0, "speech recognition and synthesis"), ("T10860", 1.0, "speech and audio processing"),115 ("T11309", 1.0, "music and audio processing"), ("T11349", 0.6, "music technology and sound studies"),116 ("T11447", 0.6, "blind source separation"), ("T12031", 0.5, "speech and dialogue systems"),117 ("T10863", 0.3, "voice and speech disorders"), ("T11034", 0.2, "digital filter design"),118 ("T10931", 0.2, "direction-of-arrival estimation (microphone arrays)"), ("T10941", 0.1, "musicology")],119 subfields=[(1711, 0.2)],120 arxiv=[("eess.AS", 1.0), ("cs.SD", 1.0), ("cs.CL", 0.1), ("eess.SP", 0.1)],121 streams=[("conf/interspeech", 1.0), ("journals/taslp", 1.0), ("journals/speech", 1.0), ("conf/slt", 1.0),122 ("conf/asru", 1.0), ("journals/csl", 0.8), ("conf/icassp", 0.5), ("conf/apsipa", 0.3), ("conf/eusipco", 0.3),123 ("journals/spl", 0.2), ("journals/sigpro", 0.2), ("journals/dsp", 0.2), ("journals/jstsp", 0.2),124 ("journals/cssp", 0.1), ("conf/mm", 0.1), ("journals/tmm", 0.1)],125 keywords=[126 (r"(?i)\bspeech (recognition|synthesis|enhancement|separation|processing|emotion|translation|coding|to text|to speech|language models?|foundation models?|representations?|technolog(y|ies)|signals?|corpus|corpora|data)\b|\btext-?to-?speech\b|\bTTS\b|\bASR\b|\bspeech-?to-?text\b", 1.0, "both"),127 (r"(?i)\b(automatic speech|spoken (language|dialogue|term|question)|speaker (recognition|verification|identification|diarization|embeddings?|adaptation)|voice (conversion|cloning|assistants?|activity detection|biometrics|synthesis)|keyword spotting|wake[- ]word|acoustic (model(s|ing)?|scene|event|features?|echo)|phonemes?|phonetic|prosod(y|ic)|intonation|lip[- ]reading|audio-?visual speech|dysarthri|pathological speech|whisper(ed)? speech)\b", 1.0, "both"),128 (r"(?i)\b(audio|music|sound|acoustic) (classification|generation|synthesis|tagging|retrieval|processing|understanding|event detection|source separation|representation|captioning|language models?|foundation models?|codecs?|signals?|features?|embeddings?)\b", 1.0, "both"),129 (r"(?i)\b(music information retrieval|MIR\b|music (transcription|recommendation|generation|genre|emotion|structure|similarity)|singing voice|beat tracking|chord (recognition|estimation)|melody extraction|symbolic music|MIDI|audio (signal|watermark|deepfake|anti-?spoofing)|microphone arrays?|beamforming for speech|hearing aids?|cochlear|room (impulse|acoustics)|reverberation|dereverberation|noise suppression|sound event|environmental sound|bird (song|call) (detection|classification)|bioacoustic|wav2vec|Whisper\b|HuBERT|WavLM|WaveNet|vocoders?|HiFi-?GAN|spectrograms?|mel[- ]spectrogram|MFCCs?|EnCodec|SoundStream)\b", 0.9, "both"),130 (r"(?i)\b(speech|audio|acoustic|speaker|music|voice|sound|auditory|phonological)\b", 0.3, "both"),131 ],132 exclusions=[],133 related=["nlp", "multimodal_ai", "generative_ai", "llm"],134 parent_of=[], child_of=[],135 provenance="OpenAlex T10201/T10860/T11309 and the eess.AS/cs.SD arXiv pair; INTERSPEECH/TASLP strong, ICASSP partial (shared with signal processing).",136),137]138