Systems and methods for reducing distortion in end-to-end feature compression in coding of multi-dimensional data
AssigneeSharp Kabushiki Kaisha
InventorsTae Meon BAE, Tianying JI, Sachin G. DESHPANDE
Life of this patent
Estimating the term…
Abstract
A device may be configured to reduce distortion in compressed feature data according to one or more of the techniques described herein. In one example, a bitstream including compressed feature data may be decoded according to the video coding standard. A quantization parameter or target bit rate and a picture type may be determined for a decoded picture corresponding to a channel. A distortion reduction engine may be selected based on the quantization parameter and the picture type. The distortion reduction engine may be applied to reduce distortion.
Claims3
A method of mitigating distortion in compressed feature data, the method comprising:
- receiving a bitstream including compressed feature data further coded according a video coding standard, wherein the compressed feature data is a tensor with channel, height, and weight dimensions;
- decoding the bitstream according to the video coding standard, such that a decoded picture corresponds to a channel;
- determining a quantization parameter and a picture type for a decoded picture;
- selecting a distortion reduction engine based on the quantization parameter and the picture type;
- and
- applying the distortion reduction engine to the decoded picture.
A device comprising one or more processors configured to:
- receive a bitstream including compressed feature data further coded according a video coding standard, wherein the compressed feature data is a tensor with channel, height, and weight dimensions;
- decode the bitstream according to the video coding standard, such that a decoded picture corresponds to a channel;
- determine a quantization parameter and a picture type for a decoded picture;
- select a distortion reduction engine based on the quantization parameter and the picture type;
- and
- apply the distortion reduction engine to the decoded picture.
Dependencies are inferred from the claim text.