Arad, BoazBen-Shahar, OhadTimofte, RaduVan Gool, LucZhang, LeiYang, Ming-HsuanXiong, ZhiweiChen, ChangShi, ZhanLiu, DongWu, FengLanaras, CharisGalliani, SilvanoSchindler, KonradStiebel, TarekKoppers, SimonSeltsam, PhilippZhou, RuofanEl Helou, MajedLahoud, FayezShahpaski, MarjanZheng, KeGao, LianruZhang, BingCui, XiminYu, HaoyangBaran Can, YigitAlvarez-Gila, AitorVan de Weijer, JoostGarrote, EstibalizGaldran, AdrianSharma, ManojKoundinya, SriharshaUpadhyay, AvinashManekar, RaunakMukhopadhyay, RudrabhaSharma, HimanshuChaudhury, SantanuNagasubramanian, KoushikGhosal, SambuddhaK. Singh, AsheeshSingh, ArtiGanapathysubramanian, BaskarSarkar, Soumik2018-11-152018-11-152018-11-15201810.1109/CVPRW.2018.00138https://infoscience.epfl.ch/handle/20.500.14299/151462WOS:000457636800131This paper reviews the first challenge on spectral image reconstruction from RGB images, i.e., the recovery of whole-scene hyperspectral (HS) information from a 3- channel RGB image. The challenge was divided into 2 tracks: the “Clean” track sought HS recovery from noise- less RGB images obtained from a known response func- tion (representing spectrally-calibrated camera) while the “Real World” track challenged participants to recover HS cubes from JPEG-compressed RGB images generated by an unknown response function. To facilitate the challenge, the BGU Hyperspectral Image Database was extended to provide participants with 256 natural HS training images, and 5+10 additional images for validation and testing, re- spectively. The “Clean” and “Real World” tracks had 73 and 63 registered participants respectively, with 12 teams competing in the final testing phase. Proposed methods and their corresponding results are reported in this review.spectral reconstructionhyperspectraltext::conference output::conference proceedings::conference paper