The Computational Imaging Lab at Portland State University develops imaging technologies that combine camera hardware, mathematical models, and computational algorithms to capture and interpret visual information.

Led by Assistant Professor Atul Ingle in the Department of Computer Science, the lab investigates imaging challenges that conventional cameras cannot easily solve, including capturing information at extremely short timescales, measuring three-dimensional environments, and performing visual analysis with limited computing power.

What We Research

Instead of treating a camera as a device that simply records an image, computational imaging considers how the sensing hardware and the algorithms that interpret its measurements can be designed together. This approach creates opportunities for imaging systems that capture different types of information, operate more efficiently, and support challenging real-world applications.

Research areas include:

  • Single-photon imaging
  • High-speed three-dimensional sensing
  • Computational camera design
  • Efficient visual processing
  • Computer vision

This work brings together optics, image processing, computer science, and machine learning.

Single-Photon Imaging

A conventional camera records light collected during an exposure. Single-photon imaging technologies can detect extremely small numbers of individual light particles, called photons. This sensitivity enables imaging research under conditions where conventional cameras may struggle to capture useful information.

The lab investigates single-photon systems capable of recording information at nanosecond and picosecond timescales. Such measurements show when light returns to a sensor, which provides information useful for measuring distance and reconstructing three-dimensional scenes.

Three-Dimensional Imaging and Depth Sensing

Three-dimensional imaging allows a system to estimate not only what objects look like, but also how far away they are. Depth information is important for applications such as robotics and autonomous vehicles, where systems must recognize and navigate their surroundings.

The lab investigates sensing methods that capture three-dimensional information rapidly and efficiently. Its research includes techniques for compressing and processing single-photon imaging measurements while preserving the information needed to reconstruct scene geometry. Examples include EDPHi, which uses equi-depth photon histograms to achieve 100x compression in single-photon 3D imaging, and DeePEDH, a learning-based method for compressive 3D captures with single-photon cameras.

More Efficient Camera Systems

Advanced cameras can produce enormous quantities of data. Transmitting and processing that information can require significant energy, memory, and computing resources. The lab studies ways to make imaging systems more efficient by reconsidering how visual data is collected and represented.

For example, the lab’s SuperCam research investigates a camera that directly captures groups of image pixels, known as superpixels, to support lightweight artificial intelligence processing. The work was published at CVPR 2026. Other projects examine compressed representations of three-dimensional imaging data. These approaches aim to reduce unnecessary processing while preserving information important for the intended application. Explore the SuperCam project

Research Applications

Computational imaging has potential applications across robotics, autonomous navigation, environmental observation, and other fields that rely on reliable sensing. By combining specialized camera technologies with computational methods, researchers can develop systems designed around particular challenges rather than depending entirely on conventional image capture.

The lab’s work contributes to the development of cameras that are faster, more sensitive, or more efficient for specific tasks. See all lab projects

Research and Learning

The Computational Imaging Lab brings together computer vision, imaging science, algorithms, and experimental hardware. Its research provides opportunities to explore both the physical principles of image formation and the computational methods used to reconstruct and interpret measurements.

The lab’s work is supported in part by grants from the National Science Foundation, the US Department of Energy, and Portland State University Venture Development Funds.