A layered light-field display is composed of several liquid crystal layers located in front of a backlight. The light rays emitted from the backlight intersect with different pixels on the layers depending on the outgoing directions. Therefore, this display can show multi-view images
(a light field) in accordance with the viewing direction. This type of displays can also be used for head-mounted displays (HMDs) thanks to its dense angular resolution. The angular resolution is an important factor, because sufficiently dense angular resolution can provide accommodation cues,
preventing visual discomfort caused by vergence accommodation conflict. To further enhance the angular resolution of a layered display, we propose to replace some of the layers with monochrome layers. While keeping the pixel size unchanged, our method can achieve three times higher resolution
than baseline architecture in the horizontal direction. To obtain a set of color and monochrome layer patterns for a target light field, we developed two computation methods based on non-negative tensor factorization and a convolutional neural network, respectively.