acldvppErase

Applicability

Description

Description: erases pixels within a specified range in the input image by using a specified value.

Prototype

Each operator has two-phase API calls. First, acldvppEraseGetWorkspaceSize is called to calculate the required workspace size based on the computation process. Then, acldvppErase is called to perform computation. The two-phase APIs are as follows:

• First-phase API:

  1	acldvppStatus acldvppEraseGetWorkspaceSize(const aclTensor *self, uint32_t top, uint32_t left, uint32_t height, uint32_t width, const aclFloatArray* value, aclTensor *out, uint64_t *workspaceSize, aclOpExecutor **executor)

• Second-phase API:

  1	acldvppStatus acldvppErase(void *workspace, uint64_t workspaceSize, aclOpExecutor *executor, aclrtStream stream)

acldvppEraseGetWorkspaceSize

• Parameters:
  • self: input tensor of the operator. Call aclCreateTensor to create data of the aclTensor type (stored on the device). The data type of the input tensor can be UINT8 or FLOAT, and the format can be NCHW or NHWC. Discontinuous tensors are not supported, in addition, N can be 1 or empty, and C can be 1 or 3 (1 indicates the input GRAY image, and 3 indicates the input RGB image).

    When the data type of a tensor is set to FLOAT, the expected value range is [0, 1].

  • top: vertical coordinate of the upper left corner of the area to be erased.
  • left: horizontal coordinate of the upper left corner of the area to be erased.
  • height: height of the area to be erased.
  • width: width of the area to be erased.
  • value: pixel filling value in the area to be erased. Call aclCreateFloatArray to create data of the aclFloatArray type. The length must be the same as the number of C channels of the input tensor. When the length is 3, the values are used to fill the R, G, and B channels, respectively.

    If a null pointer is transferred, the default value 0 is used.

    If the data type of the input tensor is FLOAT, the value range is [0, 1]. If the data type of the input tensor is UINT8, the value range is [0, 255].

  • out: operator output tensor. Call aclCreateTensor to create data of the aclTensor type (stored on the device). The data type of the output tensor can be UINT8 or FLOAT, and the format can be NCHW or NHWC. Discontinuous tensors are not supported, in addition, N can only be 1 or empty, and C can be 1 or 3 (1 indicates a GRAY image, and 3 indicates an RGB image). The values of dataType, Format, and Shape must be the same as those of self.
  • workspaceSize: size of the workspace to be allocated on the device.
  • executor: operator executor, containing the operator computation process.
• Returns:

  The status code acldvppStatus is returned. For details, see acldvpp Return Codes.

acldvppErase

• Parameters:
  • workspace: The aclrtMalloc API needs to be called to allocate the device memory. The memory size is workspaceSize. The memory address output by the aclrtMalloc API is passed here.
  • workspaceSize: The value must be the same as that obtained by calling acldvppEraseGetWorkspaceSize.
  • executor: op executor, which contains the operator computation process and is the same as the executor of the acldvppEraseGetWorkspaceSize API.
  • stream: stream for executing the task. You can reuse the created stream to save resources or call aclrtCreateStream to create a stream, and then pass the stream as an input parameter.
• Returns

  The status code acldvppStatus is returned. For details, see acldvpp Return Codes.

Constraints

The supported image resolution ranges from 6 x 4 to 4096 x 8192.

If dataType is set to UINT8, the result is the same as that of Torchvision. If dataType is set to FLOAT, the maximum difference between the result of area to be erased and Torchvision is +-1/255.