Efficient light field image transmission method based on enhanced resampling

Abstract

The invention discloses an efficient light field image transmission method based on enhanced resampling, and aims to solve the problem of large data volume in a light field image (LFI) transmission process. According to the method, a low-resolution image is generated by carrying out downsampling on a high-resolution light field image before transmission, residual information between a low-resolution upsampling (using a traditional interpolation method or a super-resolution method and the like) result and an original image is calculated, and the low-resolution image and the residual information are coded into a pseudo video sequence for transmission. After transmission, a low-resolution image is subjected to up-sampling (a traditional interpolation method or a super-resolution method and the like is used) at a decoding end, reconstruction is performed in combination with residual information, and a high-resolution light field image with the quality close to that of an original image is obtained. According to the method, the transmission data volume is reduced through downsampling, and the quality of the reconstructed image is improved by using the residual information, so that the visual effect of the image is ensured while the bandwidth requirement is reduced. The transmission method is suitable for light field image transmission, virtual reality and other application scenes needing high-quality image transmission, and efficient light field image transmission is effectively achieved.

Description

Technical Field

[0001] The present invention relates to efficient transmission of light field images (LFI), and in particular to a transmission method for reducing the amount of light field data transmission by enhancing resampling reconstruction technology and maintaining high-quality images by using residual information. The method is applicable to fields such as image processing, data compression and transmission, and is particularly applicable to virtual reality and 3D imaging application scenarios. Background Art

[0002] Light field imaging technology captures the direction and intensity of light, provides a detailed multi-dimensional information representation of the scene, and supports post-processing functions such as multi-view rendering and depth reconstruction. However, the amount of data in light field images is extremely large, especially under high-resolution conditions, which leads to high bandwidth requirements for data storage and transmission, limiting its widespread application. Traditional high-resolution light field image transmission methods usually occupy a large amount of bandwidth and storage resources because they need to transmit complete data, which greatly limits the application and promotion of light field technology. To this end, the present invention proposes an efficient transmission method based on enhanced resampling. By resampling before and after transmission and using residual information, the amount of data transmission can be effectively reduced while ensuring image quality. Summary of the invention

[0003] The present invention proposes an efficient light field image transmission method based on enhanced resampling, comprising the following steps:

[0004] Generate low-resolution light field images: Downsample the high-resolution light field images (for example, using pooling, interpolation, etc.) to generate low-resolution light field images to reduce spatial redundancy and reduce transmission bandwidth requirements.

[0005] Residual information calculation: Upsample the low-resolution image (for example, using interpolation, super-resolution, etc.), and calculate the difference between the original high-resolution image and the low-resolution upsampled image (for example, using pixel-by-pixel subtraction, etc.) to generate residual information. The residual information represents the quality difference between the upsampled image and the original image, and is used to improve the reconstruction effect after transmission.

[0006] Pseudo video sequence encoding: Low-resolution light field images and residual information are encoded into pseudo video sequences through video encoding technology (such as HEVC, JPEG, etc.). The frame rate and quantization parameters of the sequence are flexibly set according to the specific application scenario, thereby effectively compressing the light field data and reducing bandwidth requirements.

[0007] Transmit light field: Transmit the encoded pseudo video sequence in any way.

[0008] Decoding and reconstruction: Decode the pseudo video sequence at the receiving end, restore the low-resolution light field image and residual information, upsample the low-resolution image (for example, using interpolation, super-resolution, etc.), and then combine the residual information to reconstruct a high-resolution light field image that is close to the original image.

[0009] Reconstruction result output: The reconstructed high-resolution light field image is close to the original image in visual quality, achieving efficient transmission while maintaining image details and clarity. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 This is a schematic diagram of the process of efficient light field transmission based on enhanced resampling proposed by the present invention. DETAILED DESCRIPTION

[0011] Reference Figure 1 The present invention proposes an efficient light field transmission method based on enhanced resampling, comprising:

[0012] Step S1: Generate a low-resolution light field image. Downsample the high-resolution light field image to generate a low-resolution light field image, thereby reducing spatial redundancy and reducing the bandwidth requirement for data transmission. In the downsampling process, a variety of downsampling techniques such as interpolation and pooling can be used. The specific selection can be based on the actual application requirements to balance the processing speed and image quality.

[0013] Step S2: Calculate residual information. Upsample the generated low-resolution light field image and compare it with the original high-resolution image to calculate the difference and generate residual information. The residual information represents the quality difference between the upsampled image and the original image and is used to improve the image reconstruction effect after transmission. The residual information can compensate for the loss of details after the image is upsampled, thereby ensuring the quality of the reconstructed image.

[0014] Step S3: Pseudo video sequence encoding. The low-resolution light field image and residual information are encoded into a pseudo video sequence using video encoding technology. The frame rate and quantization parameters of the pseudo video sequence can be flexibly set according to the specific application scenario, thereby achieving efficient compression of the light field data and further reducing the data transmission bandwidth requirement. By encoding the light field image into a pseudo video sequence, the amount of redundant data between subviews can be effectively reduced, thereby improving transmission efficiency.

[0015] Step S4: Transmitting: The encoded pseudo video sequence is transmitted in any manner.

[0016] Step S4: Decoding and reconstruction. At the receiving end, the received pseudo video sequence is decoded to restore the low-resolution light field image and residual information. The low-resolution image is upsampled and combined with the residual information to perform an addition operation to achieve reconstruction of the high-resolution light field image. The original details of the image are restored by compensating the residual information, so that the reconstructed image is close to the original image in visual quality.

[0017] The high-resolution light field image reconstructed through the above steps is visually close to the original high-resolution image, achieving efficient transmission while maintaining image details and clarity, and is suitable for high-quality light field image transmission in low-bandwidth environments.

[0018] The method of the present invention can effectively reduce the amount of data required for light field image transmission while ensuring the quality of the image during transmission, and is particularly suitable for application scenarios such as virtual reality and 3D imaging that require high-quality light field images.

Claims

1. A light field image efficient transmission method based on enhanced resampling, characterized in that: The following steps are involved: (a) receiving a high-resolution light field image, and processing the high-resolution light field image using a downsampling technique to generate a low-resolution light field image; (b) calculating residual information between an original high-resolution light field image and an image obtained by upsampling the low-resolution light field image, and encoding the low-resolution light field image and the residual information together into a pseudo video sequence; (c) transmitting a pseudo video sequence encoded by a low-resolution light field map and residual information; (d) decoding the pseudo video sequence at a receiving end to extract a low-resolution light field image and residual information therefrom; (e) Upsampling the extracted low-resolution light field image data, and combining the residual information to obtain a high-resolution light field image with a quality close to the original image through image reconstruction technology.

2. The light field image efficient transmission method according to claim 1, characterized in that: The downsampling and upsampling steps both use traditional interpolation methods and super-resolution methods, aiming to reduce spatial redundancy and artifact effects while effectively retaining key detail information of the image.

3. The light field image efficient transmission method according to claim 1, characterized in that: The step of generating residual information adopts methods such as pixel-by-pixel subtraction.

4. The light field image efficient transmission method according to claim 1, characterized in that: In the step of generating the pseudo video sequence, one or more subviews of the light field image are encoded into a sequence of pseudo video frames using video encoding technology (such as HEVC, JPEG, etc.), thereby further reducing the amount of data transmission and improving transmission efficiency.

5. The light field image efficient transmission method according to claim 1, characterized in that: The transmission method is not limited.

6. The light field image efficient transmission method according to claim 1, characterized in that: At the decoding end, the residual information is used to enhance the reconstruction process of the light field image. By accurately compensating the difference between the upsampled image and the original high-resolution image, the accuracy and visual effect of the reconstructed light field image are significantly improved, thereby achieving high-quality image restoration.

7. The light field image efficient transmission method according to claims 1 to 4, characterized in that: The reconstruction method comprises the following steps: (a) At the decoding end, the received low-resolution light field image data is upsampled to generate a preliminary high-resolution light field image; (b) The upsampled light field image is added to its corresponding residual information to restore a high-resolution light field image with quality close to the original image.

8. The light field image efficient transmission method according to any one of claims 1 to 7, characterized in that: This method is suitable for application scenarios such as virtual reality and 3D imaging that require high-quality image transmission. It can reduce bandwidth requirements while ensuring the visual effect and transmission efficiency of the image.

Images