Sergio Goma
Bio
Sergio R. Goma, PhD, is an IEEE Fellow (Class of 2024, “for contributions to hardware implementation of image processing for color cameras in mobile phones”) and an ASIC and FPGA architect and designer with 30 years of experience across the full silicon chain, from algorithms and specification through RTL, tape-out, bring-up, and mass-production qualification. He received the B.E., M.Sc., and Ph.D. degrees in computer engineering from Universitatea Politehnica Timișoara, Romania, in 1994, 1995, and 1998. He holds more than 70 granted U.S. patents, many with European and international counterparts.
Since September 2024, Dr. Goma has focused on his own companies. He is Director of AlkemistaTech, where he designs RTL and FPGA imaging and interface cores, including ISP cores and a real-time FPGA implementation of an AI-based contrast enhancement algorithm, and contributes to the ISO/TC 42 work on depth camera performance measurement (ISO/AWI TS 8527). AlkemistaTech acquired the assets and IP of EDA4All, whose IEEE 1394b core originated with Dr. Goma; he has since updated its MIPI CSI-2 receiver core and added USB3 and PCIe interface cores. He is also a co-owner and member of Alkemista IND (DBA Sinagua Malt) in Camp Verde, Arizona, where he engineered a 20,000 sq ft malting plant that runs 24/7 with one person for operation and maintenance. He presented the concept in an invited talk at Politehnica University of Timișoara in April 2025.
From 2008 to September 2024, Dr. Goma was Senior Director of Technology at Qualcomm Technologies, where he led the Imaging Research and Development group in the chip division, including standards work. He led more than 40 engineers in San Diego, Hsinchu, and Vienna. The group completed multiple tape-outs of standalone programmable image-processing architectures and ISPs, and developed technologies in imaging, programmable architecture, and integrated optics, including 3D stacked programmable image processing with TSMC and wafer-level stacked optics with SiGe sensors. He also advised Qualcomm M&A on technical diligence and Qualcomm Ventures.
Previously, Dr. Goma was Principal Member of Technical Staff at AMD and Architect at ATI (2003 to 2008), where he was architect and lead for the image processing IP of the Imageon series of chips, which shipped more than 100 million units. He designed and implemented auto white balance, auto exposure, auto focus, color correction, and noise reduction for these chips, and was the RTL designer of the audio core in the same chip.
Dr. Goma was the technical lead for the AMD and ATI contributions to the standardization of the serial camera interface MIPI CSI-2, the most widely used camera interface in the mobile industry. His technical solutions were accepted as core technologies of both MIPI CSI-2 and MIPI DSI, including the MIPI-specific ECC code, the packet-header error handling, and the example implementation, and they became the foundation of MIPI CSI-2. He was the original RTL designer of ATI/AMD’s first MIPI cores and drove interoperability of MIPI CSI-2 across the industry.
Dr. Goma was a technical subgroup chair of CPIQ (Camera Phone Image Quality), which became IEEE 1858-2016. The work built and standardized image quality test metrics and methodologies across the industry to correlate objective test results with human perception and combine them into a meaningful consumer rating system.
Before ATI, Dr. Goma designed high-performance CCD and CMOS cameras and imaging systems for metrology and industrial automation at Photon Dynamics (now Orbotech), including very low noise metrology cameras for LCD panel testing, and at Taymer Industries.
Dr. Goma is Senior Area Editor of IEEE Transactions on Image Processing and was Associate Editor of IEEE Transactions on Computational Imaging (2015 to 2018) and the Journal of Real-Time Image Processing. He is a committee member of IS&T’s Human Vision and Electronic Imaging conference (2008 to present) and the Photography, Mobile, and Immersive Imaging conference (2007 to present), and an associate member of the IS&T Computational Imaging Technical Committee, which he served on from the start of its Special Interest Group in 2015. He is a founding member and a member of the steering committee of IS&T’s Journal of Perceptual Imaging (JPI), the peer-reviewed open-access journal that grew out of the HVEI conference. He was Vice President of the Society for Imaging Science and Technology (2014 to 2018). He received the 2014 IS&T Service Award for serving as 2014 Electronic Imaging Symposium Chair and for leading the effort for IS&T to manage the symposium on its own beginning in 2016. He was an IEEE Distinguished Industry Speaker in 2019 and is actively involved in IEEE SPS.
Dr. Goma is Finance Chair of the Computational Imaging Symposium (CISym), which he helps co-organize, and he maintains Computational Imaging Code, a curated collection of open-source code for computational imaging research. CISym grew out of the Computational Imaging (COIMG) track of IS&T Electronic Imaging, which began in 2003, and launches as an independent symposium on February 22 to 24, 2027, at the Crowne Plaza San Francisco Airport in Burlingame, with the technical co-sponsorship of the IEEE Signal Processing Society. Its program is built around talks, community-organized special sessions, and posters, and covers topics from inverse problems and AI-based imaging to sensor hardware and algorithm co-design. The organizing committee is led by Charles A. Bouman (General Chair) and Gregery T. Buzzard (Technical Chair) of Purdue University.
Dr. Goma’s research interests include programmable hardware architectures and hardware accelerators for imaging, real-time systems, computational photography, integrated imaging sensors, depth sensing, image quality and sensor metrology, image processing, and computer vision.
Lecture Topics
- Integrated Image Sensors with Compute Elements
- Programmable Architectures for Image Processing
- Algorithm-Hardware Co-Design: An Industry Perspective
- Beyond von Neumann: Compute Architectures for Real-Time Image Processing
- The One Man Factory: Concept and Implementation