High Performance Imaging

Summary: We present a novel, compact single-shot hyperspectral imaging method. It enables capturing hyperspectral images using a camera equipped with just a prism. Our method compares well with other hyperspectral imaging systems, both in terms of spectral accuracy and spatial resolution, while being orders of magnitude cheaper than commercial systems.

Author(s): Seung-Hwan Baek, Korea Advanced Institute of Science and Technology (KAIST)
Incheol Kim, Korea Advanced Institute of Science and Technology (KAIST)
Diego Gutierrez, Universidad de Zaragoza
Min H. Kim, Korea Advanced Institute of Science and Technology (KAIST)

Speaker(s): Seung-Hwan Baek, KAIST

Summary: We present a novel hyperspectral reconstruction algorithm to overcome the tradeoff between spectral and spatial accuracy. Our technique can be applied to any compressive hyperspectral imagers; it compares well with other state-of-the-art methods. We validate the performance of our algorithm on a real imaging system.

Author(s): Inchang Choi, Korea Advanced Institute of Science and Technology (KAIST)
Daniel S. Jeon, Korea Advanced Institute of Science and Technology (KAIST)
Giljoo Nam, Korea Advanced Institute of Science and Technology (KAIST)
Diego Gutierrez, Universidad de Zaragoza
Min H. Kim, Korea Advanced Institute of Science and Technology (KAIST)

Speaker(s): Inchang Choi, KAIST

Summary: We present a novel realtime approach to correct depth errors due to multipath interference in off-the-shelf Time-of-Flight cameras. We leverage the convolutional nature of light transport, and learn depth corrections using a CNN trained in a novel two-stage approach, using both real and synthetic data.

Author(s): Julio Marco, University of Zaragoza
Quercus Manuel Hernandez, Universidad de Zaragoza
Adolfo Muñoz, Universidad de Zaragoza
Yue Dong, Microsoft Research Asia
Adrian Jarabo, Universidad de Zaragoza
Min Kim, KAIST
Xin Tong, Microsoft Research Asia
Diego Gutierrez, Universidad de Zaragoza

Speaker(s): Julio Marco, Universidad de Zaragoza

Summary: A single-shot difference imaging approach is directly implemented in the sensor hardware of emerging time-of-flight cameras. We demonstrate that the proposed technique is useful for direct-global illumination separation, for direct imaging of spatial and temporal image gradients, for direct depth edge imaging, and more.

Author(s): Clara Callenberg, University of Bonn
Felix Heide, Stanford University
Gordon Wetzstein, Stanford University
Matthias Hullin, University of Bonn

Speaker(s): Clara Callenberg, University of Bonn