Function: dvpp_jpeg_decode_async

Applicability

Product	Supported
Atlas 350 Accelerator Card	x
Atlas A3 training product / Atlas A3 inference product	√
Atlas A2 training product / Atlas A2 inference product	√
Atlas training product	√
Atlas inference product	√
Atlas 200I/500 A2 inference product	√

Description

Decodes a .jpg, .jpeg, .JPG, or .JPEG image. This API is asynchronous.

Prototype

C Prototype

        
             aclError acldvppJpegDecodeAsync(acldvppChannelDesc *channelDesc, const void *data, uint32_t size, acldvppPicDesc *outputDesc, aclrtStream stream)

Python Function

        
             ret = acl.media.dvpp_jpeg_decode_async(dvpp_channel_desc, data, size, output_dvpp_pic_desc, stream)

Parameters

Parameter	Description
dvpp_channel_desc	Int, pointer address of the channel description. The value must be the same as that of dvpp_channel_desc specified when a channel is created by calling acl.media.dvpp_create_channel.
data	Int, pointer address of the memory address of the input image.
size	Int, actual size of the input image, in bytes.
output_dvpp_pic_desc	Int, pointer address of the output image information. Call the acl.media.dvpp_create_pic_desc API to create the image description. Call the acl.media.dvpp_set_pic_desc APIs to set the memory address, memory size, image format, width_stride, and height_stride of the output image. Call acl.media.dvpp_jpeg_predict_dec_size to predict the output buffer size in advance. Call acl.media.dvpp_jpeg_get_image_info/acl.media.dvpp_jpeg_get_image_info_v2 to obtain the width and height of the source JPEG image. For details about the output format and width and height alignment requirements, see Functions and Restrictions. When this argument is used as an output, the user can obtain the decoding output from the buffer address and call the acl.media.dvpp_get_pic_desc APIs to obtain the width and height of the output image. NOTE: Atlas inference product : If the decoding output will be further processed (for example, cropped and resized by VPC), use the acl.media.dvpp_get_pic_desc API to obtain the width and height of the output image, in case that the image width and height have been adjusted due to alignment requirements. For example, for JPEG (444) to YUV420SP decoding, if the input width and height are odd numbers, the output width and height are theoretically odd numbers. However, JPEGD will round down the output width and height to multiples of 2, as required by the YUV420SP format. Atlas 200I/500 A2 inference product : If the decoding output will be further processed (for example, cropped and resized by VPC), use the acl.media.dvpp_get_pic_desc API to obtain the width and height of the output image, in case that the image width and height have been adjusted due to alignment requirements. For example, for JPEG (444) to YUV420SP decoding, if the input width and height are odd numbers, the output width and height are theoretically odd numbers. However, JPEGD will round down the output width and height to multiples of 2, as required by the YUV420SP format. Atlas A2 training product / Atlas A2 inference product : If the decoding output will be further processed (for example, cropped and resized by VPC), use the acl.media.dvpp_get_pic_desc API to obtain the width and height of the output image, in case that the image width and height have been adjusted due to alignment requirements. For example, for JPEG (444) to YUV420SP decoding, if the input width and height are odd numbers, the output width and height are theoretically odd numbers. However, JPEGD will round down the output width and height to multiples of 2, as required by the YUV420SP format. Atlas A3 training product / Atlas A3 inference product : If the decoding output will be further processed (for example, cropped and resized by VPC), use the acl.media.dvpp_get_pic_desc API to obtain the width and height of the output image, in case that the image width and height have been adjusted due to alignment requirements. For example, for JPEG (444) to YUV420SP decoding, if the input width and height are odd numbers, the output width and height are theoretically odd numbers. However, JPEGD will round down the output width and height to multiples of 2, as required by the YUV420SP format.
stream	Int, stream object.

Return Value

Return Value	Description
ret	Int, error code. 0 on success; else, failure.

Restrictions

This API is asynchronous. The API call delivers a task rather than executes a task. After this API is called, call the synchronization API (for example, acl.rt.synchronize_stream) to ensure that the task is complete.

Reference

For details about the API call sequence and examples, see JPEGD.

Parent topic: JPEGD