{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/seeing-into-darkness-scotopic-visual","title":"Seeing into Darkness: Scotopic Visual Recognition","arxiv_id":"1610.00405","date":"2016-10-03","proceeding":"CVPR 2017 7","authors":["Bo Chen","Pietro Perona"],"abstract":"Images are formed by counting how many photons traveling from a given set of\ndirections hit an image sensor during a given time interval. When photons are\nfew and far in between, the concept of `image' breaks down and it is best to\nconsider directly the flow of photons. Computer vision in this regime, which we\ncall `scotopic', is radically different from the classical image-based paradigm\nin that visual computations (classification, control, search) have to take\nplace while the stream of photons is captured and decisions may be taken as\nsoon as enough information is available. The scotopic regime is important for\nbiomedical imaging, security, astronomy and many other fields. Here we develop\na framework that allows a machine to classify objects with as few photons as\npossible, while maintaining the error rate below an acceptable threshold. A\ndynamic and asymptotically optimal speed-accuracy tradeoff is a key feature of\nthis framework. We propose and study an algorithm to optimize the tradeoff of a\nconvolutional network directly from lowlight images and evaluate on simulated\nimages from standard datasets. Surprisingly, scotopic systems can achieve\ncomparable classification performance as traditional vision systems while using\nless than 0.1% of the photons in a conventional image. In addition, we\ndemonstrate that our algorithms work even when the illuminance of the\nenvironment is unknown and varying. Last, we outline a spiking neural network\ncoupled with photon-counting sensors as a power-efficient hardware realization\nof scotopic algorithms.","url_abs":"http://arxiv.org/abs/1610.00405v1","url_pdf":"http://arxiv.org/pdf/1610.00405v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"seeing-into-darkness-scotopic-visual","repo_url":"https://github.com/bochencaltech/scotopic","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":{"status":"ok"}}],"tasks":[{"task_slug":"astronomy","task_name":"Astronomy"},{"task_slug":"classification","task_name":"General Classification"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1610.00405","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}