It is important to note that modern deep learning techniques generally use more sophisticated network architecture and training strategies, better algorithms, more powerful computers and a lot more input variables, but the fundamental problems we are facing are pretty much the same as they were twenty years ago. … More Neural Network Calibrations, Old and New

Deep learning is a machine learning technique based on learning data representations, which is similar to how the human brain processes information. The sensory inputs of a human, such as visual, auditory, olfactory (smell), taste, touch, position, heat, etc., feed into a myriad of interconnected neurons that produce perception, recognition and knowledge, i.e. intelligence. In … More Using Deep Learning and Artificial Intelligence to Interpret Hyperspectral Data

  The performance of sorting equipment is characterized by sorting performance i.e. how well does the sorting machine separate the target materials, grains from debris, damaged apples from good apples or intact perfect almonds from hulls, twigs and stones that have the same size and color as the good almonds. The other important performance parameter … More High speed sorting using hyperspectral imaging

    A CMOS sensor with 512 spatial pixels and 220 spectral channels supports 330 fps at full resolution. The sensor has single or multiple ROI capability to increase frame rates past 6000 fps.   SPECIM has always been very strong with high performance imaging spectrographs. The FX 10 goes to a new level featuring … More NEW FROM SPECIM SPECTRAL IMAGING

This year’s $1 million TED prize winner, Sarah Parcak, space archaeologist, was recently explaining spectral imaging and infrared imaging to Stephen Colbert. Her prize winning talk included an explanation of how she extracted archeological information, outlines of undiscovered ruins, pyramids and towns, from spectral changes in high-resolution NASA satellite imagery. Data treatments familiar to practitioners of … More Satellite Archeology using spectral imaging

Hyperspectral imaging applications produce massive amounts of data which require the right software tools for analysis. Before digging into the complex algorithms and chemometric analyses, however, it is useful to visualize and explore the data a bit. At MSV, we do sample measurements for a wide range of applications to check out whether a problem might be … More Hyperspectral Image Visualization Software