An optimization method for online conference system

Abstract

This invention discloses an optimization method for an online conferencing system, relating to the field of online conferencing technology. The method includes: optimization of image processing and transmission: optimizing image encoding and decoding algorithms; bandwidth optimization: automatically adjusting the image transmission bitrate using adaptive bitrate technology based on the user's device and network conditions; optimization of user interaction: adding functions such as liking, sharing, real-time voice calls, and real-time video calls; and optimization of device compatibility: automatically adjusting the image display resolution using a responsive design adapted to different devices and screen resolutions based on the user's device and network conditions. This invention's optimization method for the online conferencing system can effectively reduce image processing and transmission latency and improve image transmission quality, while reducing bandwidth requirements, enriching user interaction functions, and improving device compatibility.

Description

Technical Field

[0001] This invention relates to the field of online conferencing technology, and in particular to an optimization method and system for an online conferencing system. Background Art

[0002] Today, with the rapid development of technologies such as mobile devices, high-speed networks, and cloud computing, holding meetings and conferences online can effectively save time and manpower resources, while also offering the advantages of convenience, speed, and efficiency.

[0003] Existing technology involves providing online meeting systems for users to participate in meetings and conferences. However, existing online meeting systems still have some technical shortcomings, such as poor image transmission quality, high bandwidth requirements, insufficient functionality leading to poor user interactivity, and poor device compatibility.

[0004] Therefore, in order to address the shortcomings of existing online meeting systems, this invention proposes an optimization method for online meeting systems. Summary of the Invention

[0005] The present invention proposes an optimization method for an online meeting system, which can effectively solve the technical problems of poor image transmission quality, high bandwidth requirements, poor user interactivity, and poor device compatibility in existing online meeting systems.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] An optimization method for an online meeting system includes:

[0008] Optimization of image processing and transmission: Optimizing image encoding and decoding algorithms;

[0009] Bandwidth optimization: The image transmission bitrate is automatically adjusted using adaptive bitrate technology based on the user's device and network conditions;

[0010] Optimizations for user interaction: Added features such as likes, sharing, real-time voice calls, and real-time video calls;

[0011] Device compatibility optimization: Based on the user's device and network conditions, the image display resolution is automatically adjusted using a responsive design that adapts to different devices and screen resolutions.

[0012] Optionally, the algorithm for optimizing image encoding and decoding specifically includes:

[0013] Image encoding process: The image is sequentially processed by color space conversion, block division, DCT transformation, quantization, zigzag sorting, DC differential coding, Huffman coding, and channel coding;

[0014] Image decoding process: The image after the encoding process is sequentially processed by channel decoding, Huffman decoding, Z-shaped inverse sorting and DC inverse difference, inverse quantization, inverse DCT transform and color space restoration.

[0015] Optionally, the optimization of image processing and transmission may further include: optimizing image transmission protocols, optimizing image compression algorithms, using hardware acceleration, optimizing the network environment, adopting preprocessing and postprocessing strategies, and optimizing servers.

[0016] Optionally, the optimization of image processing and transmission further includes:

[0017] Training deep learning models for image enhancement and restoration;

[0018] The trained deep learning model is used to enhance and repair the image in real time.

[0019] Optionally, the specific steps for training the deep learning model for image enhancement and restoration include:

[0020] Data preparation steps: Collect image data and preprocess it to serve as training sample set, test sample set and validation sample set respectively;

[0021] Model selection and design steps: Select or design a deep learning model suitable for image enhancement and restoration;

[0022] Model training steps: Train the deep learning model using the training sample set;

[0023] Model testing and validation steps: Use the test sample set and the validation sample set to test and validate the capabilities of the trained deep learning model, respectively.

[0024] Optionally, the preprocessing includes size normalization, pixel normalization, and data augmentation.

[0025] Optionally, the selection or design of a deep learning model suitable for image enhancement and restoration includes selecting a convolutional neural network or a generative adversarial network.

[0026] Optionally, the real-time enhancement and restoration of the image using a trained deep learning model includes:

[0027] The image is input into the trained deep learning model to obtain the enhanced or repaired image output by the deep learning model. The output image is evaluated, and the output image is manually processed based on the evaluation results.

[0028] Optionally, evaluating the output image includes assessing the image quality using image quality metrics such as PSNR and SSIM.

[0029] Optionally, the bandwidth optimization further includes: using CDN technology to cache the image on a distributed edge server.

[0030] Compared with the prior art, the beneficial effects of this invention are as follows:

[0031] By simultaneously optimizing image processing and transmission, bandwidth, user interaction, and device compatibility, it is possible to effectively reduce the latency of image processing and transmission in online conferencing systems, improve image transmission quality, lower bandwidth requirements, enrich user interaction functions, and enhance device compatibility. Attached Figure Description

[0032] Figure 1 This is a schematic diagram illustrating an optimized method for online meetings provided by the present invention. Detailed Implementation

[0033] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0034] Please see Figure 1 The diagram shown illustrates an optimization method for an online meeting system provided by the present invention, comprising the following steps:

[0035] Optimization of image processing and transmission: Optimizing image encoding and decoding algorithms;

[0036] Bandwidth optimization: The image transmission bitrate is automatically adjusted using adaptive bitrate technology based on the user's device and network conditions;

[0037] Optimizations for user interaction: Added features such as likes, sharing, real-time voice calls, and real-time video calls;

[0038] Device compatibility optimization: Based on the user's device and network conditions, the image display resolution is automatically adjusted using a responsive design that adapts to different devices and screen resolutions.

[0039] Specifically, by optimizing the algorithms or processes of image encoding and image decoding, the latency in image processing and transmission can be effectively reduced, while ensuring the quality of image transmission, thus achieving optimization of image processing and transmission.

[0040] By monitoring the status of the user's device and network in real time, and then automatically adjusting the image transmission bitrate based on the real-time monitoring results using adaptive bitrate technology, the image transmission is adapted to the user's current bandwidth status as much as possible, avoiding excessive latency and thus optimizing bandwidth.

[0041] By adding features that effectively enhance interactivity, such as likes, sharing, real-time voice calls, and real-time video calls, user interaction can be optimized. In other embodiments, other interactive features can be added.

[0042] By monitoring the status of users' devices and networks in real time, and then providing users with responsive designs adapted to different devices and screen resolutions based on the real-time monitoring results, the system can automatically adjust the image display resolution on each user's device to ensure that the user's device can be adapted to the image display size, thereby optimizing device compatibility.

[0043] Therefore, by simultaneously optimizing aspects such as image processing and transmission, bandwidth, user interaction, and device compatibility, it is possible to effectively reduce latency during image processing and transmission, lower bandwidth requirements, enrich user interaction features, and improve device compatibility.

[0044] Adaptive Bitrate Streaming (ABR) is a streaming technology that adjusts the quality of the video image in real time based on the viewer's network conditions. The main steps of this technology include:

[0045] Content preparation: The video is pre-encoded into multiple versions at different bitrates and cut into small segments by time. Each segment of each version is stored and indexed.

[0046] Connection establishment: When a client requests playback, the server first checks the client's network conditions, including bandwidth and latency, and then selects the appropriate version for playback based on these conditions.

[0047] Continuous monitoring: During playback, the client continuously monitors the network status and the player buffer status. When a change in bandwidth is detected, or the buffer data falls below a certain threshold, the client switches to the appropriate version for playback.

[0048] Version switching: When a version switch is required, the client requests the new version when retrieving the next or subsequent segments. Each version of the video contains the same content, only the bitrate is different, so a seamless switching can be achieved for the user.

[0049] The key to adaptive bitrate technology is to achieve real-time and continuous monitoring of network conditions and seamless switching between versions. This may involve issues such as bandwidth estimation, version selection strategies, and switching strategies. Furthermore, adaptive bitrate technology has different implementation methods, and therefore may correspond to different strategies and algorithms.

[0050] In one optional embodiment, the algorithm for optimizing image encoding and decoding specifically includes:

[0051] Image encoding process: The image is processed sequentially by color space conversion, block division, DCT transformation, quantization, zigzag sorting, DC differential coding, Huffman coding, and channel coding;

[0052] Image decoding process: The image after encoding is sequentially processed by channel decoding, Huffman decoding, Z-shaped inverse sorting and DC inverse difference, inverse quantization, inverse DCT transform, and color space restoration.

[0053] Specifically, the image encoding process is the image compression process. By encoding images, storage space can be effectively saved and processing and transmission time and costs can be reduced, thereby helping to reduce the latency generated during image processing and transmission.

[0054] Color space conversion of an image refers to converting its color model, specifically from the RGB color model to the YCbCr color model. The RGB color model, also known as the red-green-blue color model, uses the three primary colors of light—red, green, and blue—added in different proportions to produce a variety of colors. In YCbCr, Y represents the luminance component, Cb represents the blue chromaticity component, and Cr represents the red chromaticity component. The human eye is more sensitive to the Y component of video; therefore, after subsampling the chromaticity component to reduce its intensity, the change in image quality is imperceptible to the naked eye. Thus, converting the image's color model from RGB to YCbCr helps to separate color and luminance information for processing, thereby increasing and enhancing the efficiency and effectiveness of image compression.

[0055] Segmentation refers to dividing an image converted to the YCbCr color model into multiple smaller blocks, specifically 8x8 blocks.

[0056] DCT, or Discrete Cosine Transform, is a mathematical transformation that transforms an image from the spatial domain to the frequency domain by performing DCT on the data of each image block after it has been divided into blocks, making it easier to compress the image further.

[0057] Quantization refers to the process of quantizing the DCT transform result using a quantization table. The degree of quantization can be adjusted according to the requirements of image quality and compression ratio. Quantization in the frequency domain after DCT transform involves rounding the data by a certain step size to reduce data precision.

[0058] Z-sorting and DC differential coding refer to converting quantized two-dimensional data into one-dimensional data and simultaneously performing differential coding on the DC coefficients.

[0059] Huffman coding is a type of entropy coding. It encodes the data obtained in the previous step using Huffman coding or arithmetic coding, representing frequently occurring patterns with shorter codewords and infrequently occurring patterns with longer codewords, which helps to further reduce the amount of data.

[0060] Channel coding refers to encoding based on the characteristics of the data transmission medium, which helps to increase the error correction and detection capabilities of the data.

[0061] After all the above encoding processes, encapsulation is usually required, which means encapsulating the compressed data and some necessary header information, metadata, etc., into a specific image file format, such as JPEG format, PNG format, etc.

[0062] The channel decoding, Huffman decoding, Z-ordering and DC inverse differential, inverse quantization, inverse DCT transform, and color space restoration in the decoding process are the reverse processing operations of the channel encoding, Huffman encoding, Z-ordering and DC differential encoding, quantization, DCT transform, block division, and color space conversion in the encoding process.

[0063] Through the above-mentioned multiple encoding and decoding steps, image data can be fully compressed and decompressed, which helps to improve the speed of image transmission and processing while effectively ensuring the quality of the decompressed or restored image.

[0064] Compressing images can reduce file size, thus decreasing transmission time. However, excessive compression should be avoided to prevent image quality degradation.

[0065] In the actual implementation process, image encoding and decoding involve many complex mathematical operations and encoding theories. It is necessary to select appropriate algorithms and parameters according to the requirements of the actual application scenario, and adjust the optimization strategy accordingly.

[0066] In one alternative embodiment, the optimization of image processing and transmission further includes: optimizing the image transmission protocol, optimizing the image processing algorithm, using hardware acceleration, optimizing the network environment, adopting preprocessing and postprocessing strategies, and optimizing the server.

[0067] Specifically, in the process of image processing and transmission, latency can also be reduced by optimizing the entire process.

[0068] Optimizing the image transmission protocol means choosing a transmission protocol with higher real-time requirements, such as using UDP instead of TCP.

[0069] Optimizing image processing algorithms specifically includes reducing computational load, selecting better-performing algorithms, and providing parallel processing mechanisms, thereby effectively saving image processing time and improving image processing efficiency.

[0070] Hardware acceleration refers to using dedicated hardware such as a GPU to accelerate processing, or using hardware codecs to encode and decode images.

[0071] Optimizing the network environment means ensuring sufficient network bandwidth and low network latency to ensure real-time, fast image transmission.

[0072] Employing preprocessing and postprocessing strategies allows for greater compression ratios on non-critical frames and skipping certain frames when the network is busy, effectively reducing image transmission latency.

[0073] Optimizing a server includes optimizing server load, queuing strategies, and load balancing, which helps improve the server's latency handling or enables point-to-point real-time communication between the server and the client based on technologies such as WebRTC, reducing latency caused by server transmission.

[0074] Specifically, the aforementioned optimization methods for image processing and transmission need to be selected and adjusted according to local conditions, and sufficient testing is required to ensure that they do not adversely affect the image quality in order to find the most suitable and balanced optimization scheme.

[0075] In an alternative embodiment, the optimization of image processing and transmission further includes:

[0076] Training deep learning models for image enhancement and restoration;

[0077] Use a trained deep learning model to enhance and repair images in real time.

[0078] Specifically, by combining artificial intelligence and image processing technologies, the AI ​​model can perform deep learning to enhance and repair images in real time, which is beneficial to further improve the quality of image transmission.

[0079] In one alternative embodiment, the specific steps for training a deep learning model for image enhancement and restoration include:

[0080] Data preparation steps: Collect image data and preprocess it to serve as training sample set, test sample set and validation sample set respectively;

[0081] Model selection and design steps: Select or design a deep learning model suitable for image enhancement and restoration;

[0082] Model training steps: Train the deep learning model using the training sample set;

[0083] Model testing and validation steps: Use test and validation sample sets to test and validate the capabilities of the trained deep learning model, respectively.

[0084] The collected image data can include a large number of raw images or enhanced and repaired images. Preprocessing includes size normalization, pixel normalization, and data augmentation, which helps to improve the quality of the collected image data.

[0085] Deep learning models suitable for image enhancement and restoration include those commonly used in image restoration tasks, such as convolutional neural networks or generative adversarial networks.

[0086] Specifically, the model training process typically involves forward propagation and backward propagation, and the model parameters are continuously adjusted to minimize model error.

[0087] The model testing and validation steps involve testing and validating the ability of the trained deep learning model to generalize, thus ensuring the model's effectiveness.

[0088] In one alternative embodiment, real-time enhancement and restoration of images using a trained deep learning model includes:

[0089] The image is input into the trained deep learning model to obtain the enhanced or repaired image output by the deep learning model. The output image is evaluated and processed according to the evaluation results.

[0090] Specifically, after a deep learning model is trained, when a new image that needs to be enhanced or repaired is input into the model, the model will automatically enhance or repair the new image and then output the enhanced or repaired new image.

[0091] In one alternative embodiment, after the image is output, quantitative indicators such as PSNR and SSIM can be used to evaluate the image. In other embodiments, a comprehensive manual evaluation can also be performed.

[0092] Among them, PSNR (Peak Signal-to-Noise Ratio) and SSIM (Structural Similarity) are image quality evaluation metrics. Other image quality metrics that can be used in this embodiment include LPIPS and MSE. By adopting these image quality evaluation metrics and combining various evaluation metrics, a more scientific, balanced, and reasonable evaluation result can be obtained.

[0093] After obtaining the evaluation results, the output image can be processed, including further enhancement and restoration of the image by manual or other methods.

[0094] The above steps roughly describe the process of image enhancement and restoration using artificial intelligence and image processing technologies. The actual steps may vary and need to be adjusted according to specific task requirements and environment.

[0095] In an alternative embodiment, bandwidth optimization further includes caching images on distributed edge servers using CDN technology.

[0096] Specifically, using technologies such as CDN (Content Delivery Network) to cache images on distributed edge servers can effectively reduce server load and improve image transmission efficiency.

[0097] In summary, the technical solution of this application, through various optimizations in image processing and transmission, bandwidth, user interaction, and device compatibility, can effectively reduce the latency of image processing and transmission, thereby ensuring image transmission speed and quality. It is also conducive to adapting to various bandwidth conditions, enriching user interaction functions, and adapting to various devices, thus improving the user convenience and satisfaction of online meetings and enabling such online meetings to fully adapt to and meet the needs of various users.

[0098] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. An optimization method for online meetings, characterized in that, include: Optimization of image processing and transmission: Optimizing image encoding and decoding algorithms; Bandwidth optimization: The image transmission bitrate is automatically adjusted using adaptive bitrate technology based on the user's device and network conditions; Optimizations for user interaction: Added features such as likes, sharing, real-time voice calls, and real-time video calls; Optimization for device compatibility: Based on the user's device and network conditions, the responsive design automatically adjusts the resolution of the image display to suit different devices and screen resolutions; The algorithm for optimizing image encoding and decoding specifically includes: Image encoding process: The image is sequentially processed by color space conversion, block division, DCT transformation, quantization, zigzag sorting, DC differential coding, Huffman coding, and channel coding; Image decoding process: The image after the encoding process is sequentially processed by channel decoding, Huffman decoding, Z-shaped inverse sorting and DC inverse difference, inverse quantization, inverse DCT transform, and color space restoration; the optimization of image processing and transmission also includes: optimizing image transmission protocols, optimizing image compression algorithms, using hardware acceleration, optimizing network environment, adopting preprocessing and post-processing strategies, and optimizing servers; the optimization of image processing and transmission also includes: Training deep learning models for image enhancement and restoration; The image is enhanced and repaired in real time using a trained deep learning model; the specific steps for training the deep learning model for image enhancement and repair include: Data preparation steps: Collect image data and preprocess it to serve as training sample set, test sample set and validation sample set respectively; Model selection and design steps: Select or design a deep learning model suitable for image enhancement and restoration; Model training steps: Train the deep learning model using the training sample set; Model testing and validation steps: The capabilities of the trained deep learning model are tested and validated using the test sample set and the validation sample set, respectively; the preprocessing includes size normalization, pixel normalization, and data augmentation; the selection or design of a suitable deep learning model for image enhancement and restoration includes selecting a convolutional neural network or a generative adversarial network; the real-time enhancement and restoration of the image using the trained deep learning model includes inputting the image into the trained deep learning model, obtaining the enhanced or restored image output by the deep learning model, evaluating the output image, and manually processing the output image based on the evaluation results; the evaluation of the output image includes evaluating the image quality using image quality evaluation metrics including PSNR and SSIM; the bandwidth optimization further includes caching the image on a distributed edge server using CDN technology.

Images