{"url":"/method/spatial-feature-transform","slug":"spatial-feature-transform","name":"Spatial Feature Transform","full_name":"Spatial Feature Transform","full_name_withheld":false,"description_markdown":"**Spatial Feature Transform**, or **SFT**, is a layer that generates affine transformation parameters for spatial-wise feature modulation, and was originally proposed within the context of image super-resolution. A Spatial Feature Transform (SFT) layer learns a mapping function $\\mathcal{M}$ that outputs a modulation parameter pair $(\\mathbf{\\gamma}, \\mathbf{\\beta})$ based on some prior condition $\\Psi$. The learned parameter pair adaptively influences the outputs by applying an affine transformation spatially to each intermediate feature maps in an SR network. During testing, only a single forward pass is needed to generate the HR image given the LR input and segmentation probability maps.\r\n\r\nMore precisely, the prior $\\Psi$ is modeled by a pair of affine transformation parameters $(\\mathbf{\\gamma}, \\mathbf{\\beta})$ through a mapping function $\\mathcal{M}: \\Psi \\mapsto(\\mathbf{\\gamma}, \\mathbf{\\beta})$. Consequently,\r\n\r\n$$\r\n\\hat{\\mathbf{y}}=G_{\\mathbf{\\theta}}(\\mathbf{x} \\mid \\mathbf{\\gamma}, \\mathbf{\\beta}), \\quad(\\mathbf{\\gamma}, \\mathbf{\\beta})=\\mathcal{M}(\\Psi)\r\n$$\r\n\r\nAfter obtaining $(\\mathbf{\\gamma}, \\mathbf{\\beta})$ from conditions, the transformation is carried out by scaling and shifting feature maps of a specific layer:\r\n\r\n$$\r\n\\operatorname{SFT}(\\mathbf{F} \\mid \\mathbf{\\gamma}, \\mathbf{\\beta})=\\mathbf{\\gamma} \\odot \\mathbf{F}+\\mathbf{\\beta}\r\n$$\r\n\r\nwhere $\\mathbf{F}$ denotes the feature maps, whose dimension is the same as $\\gamma$ and $\\mathbf{\\beta}$, and $\\odot$ is referred to element-wise multiplication, i.e., Hadamard product. Since the spatial dimensions are preserved, the SFT layer not only performs feature-wise manipulation but also spatial-wise transformation.","description_state":"present","introduced_year":null,"introduced_by":{"title":"Recovering Realistic Texture in Image Super-resolution by Deep Spatial Feature Transform","paper":"/paper/recovering-realistic-texture-in-image-super","first_author":"Xintao Wang","n_authors":4,"url_abs":null,"archive_paper_url":"https://paperswithcode.com/paper/recovering-realistic-texture-in-image-super"},"source":{"url":"http://arxiv.org/abs/1804.02815v1","title":"Recovering Realistic Texture in Image Super-resolution by Deep Spatial Feature Transform","url_on_a_paper_host":true},"code_snippet_url":null,"code_snippet_url_on_a_code_host":false,"categories":[{"area":"Computer Vision","area_id":"computer-vision","collection":"Image Model Blocks","url":"/methods/category/image-model-blocks","pwc_aliases":[]}],"n_papers_tagged":8,"archive_num_papers":8,"papers_newest_first":[{"paper":"/paper/jafar-jack-up-any-feature-at-any-resolution-1","title":"JAFAR: Jack up Any Feature at Any Resolution","date":"2025-06-10","arxiv_id":"2506.11136","n_code_links":1,"syntology":null},{"paper":null,"title":"HeightLane: BEV Heightmap guided 3D Lane Detection","date":"2024-08-15","arxiv_id":"2408.08270","n_code_links":0,"syntology":null},{"paper":"/paper/arc2face-a-foundation-model-of-human-faces","title":"Arc2Face: A Foundation Model for ID-Consistent Human Faces","date":"2024-03-18","arxiv_id":"2403.11641","n_code_links":3,"syntology":{"ran":1,"of":4,"unverified":3,"pointer_only":3}},{"paper":null,"title":"Boosting Cross-Quality Face Verification using Blind Face Restoration","date":"2023-08-15","arxiv_id":"2308.07967","n_code_links":0,"syntology":null},{"paper":"/paper/variable-rate-deep-image-compression-through","title":"Variable-Rate Deep Image Compression through Spatially-Adaptive Feature Transform","date":"2021-08-21","arxiv_id":"2108.09551","n_code_links":1,"syntology":{"ran":14,"of":23,"unverified":9,"pointer_only":23}},{"paper":"/paper/towards-real-world-blind-face-restoration","title":"Towards Real-World Blind Face Restoration with Generative Facial Prior","date":"2021-01-11","arxiv_id":"2101.04061","n_code_links":1,"syntology":null},{"paper":"/paper/blind-super-resolution-with-iterative-kernel","title":"Blind Super-Resolution With Iterative Kernel Correction","date":"2019-04-06","arxiv_id":"1904.03377","n_code_links":3,"syntology":{"ran":2,"of":6,"unverified":4,"pointer_only":0}},{"paper":"/paper/recovering-realistic-texture-in-image-super","title":"Recovering Realistic Texture in Image Super-resolution by Deep Spatial Feature Transform","date":"2018-04-09","arxiv_id":"1804.02815","n_code_links":4,"syntology":null}],"papers_shown":8,"tasks":[{"task":"/task/blind-face-restoration","name":"Blind Face Restoration","papers":2},{"task":"/task/face-recognition","name":"Face Recognition","papers":2},{"task":"/task/image-super-resolution","name":"Image Super-Resolution","papers":2},{"task":"/task/super-resolution","name":"Super-Resolution","papers":2},{"task":"/task/3d-lane-detection","name":"3D Lane Detection","papers":1},{"task":"/task/blind-super-resolution","name":"Blind Super-Resolution","papers":1},{"task":"/task/diffusion-personalization","name":"Diffusion Personalization","papers":1},{"task":"/task/diffusion-personalization-tuning-free","name":"Diffusion Personalization Tuning Free","papers":1},{"task":"/task/face-generation","name":"Face Generation","papers":1},{"task":"/task/face-verification","name":"Face Verification","papers":1},{"task":"/task/feature-upsampling","name":"Feature Upsampling","papers":1},{"task":"/task/image-classification","name":"Image Classification","papers":1},{"task":"/task/image-compression","name":"Image Compression","papers":1},{"task":"/task/lane-detection","name":"Lane Detection","papers":1},{"task":"/task/semantic-segmentation","name":"Semantic Segmentation","papers":1},{"task":"/task/video-super-resolution","name":"Video Super-Resolution","papers":1},{"task":"/task/image-classification","name":"image-classification","papers":1}],"tasks_shown":17,"n_tasks":17,"usage_by_year":[{"year":"2018","papers":1},{"year":"2019","papers":1},{"year":"2021","papers":2},{"year":"2023","papers":1},{"year":"2024","papers":2},{"year":"2025","papers":1}],"row_source":"methods_table","archive":{"source":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","archive_url":"https://paperswithcode.com/method/spatial-feature-transform"},"syntology_read_at":"2026-09-24T18:15:14+00:00"}