# seefine-api-img-v1.remove-background ## 1 Short Description These APIs employ highly efficient and accurate algorithms to swiftly, automatically, and in bulk, remove backgrounds or target objects from images. ### Removing Backgrounds or Objects The user has the option to remove either the background or a specific target object. Upon completion, the system provides both the modified image, devoid of the background or object, and a masked image representing the removed area, facilitating further processing of the image or the removed region. For the image with the background or object removed, the user can select to fill the void space with either white or black. Additionally, the user can opt to receive the image in PNG format, which includes a transparent layer (potentially resulting in a larger file size), or in JPG format, which offers a smaller file size and faster processing times. The option to simultaneously receive the masked image of the removed area is also available, with the masked image being provided in JPG format. ### Removal and Impainting When removing a target object, the user can leverage APIs that incorporate impainting capabilities, enabling the automatic restoration of the image area where the object was previously removed, resulting in a seamless and complete output image. These APIs can be considered as a form of advanced image erasure tool. Furthermore, the user retains the option to receive the masked image of the removed area, facilitating additional post-processing or analysis. ### Removing Anything with Prompt and Impainting The user can also employ textual descriptions to identify and remove target objects from images. Provided the object within the image can be linguistically described, it can be excised from the image. By selecting APIs that support both "remove anything" and "impaint" functionalities, the user can not only remove any object but also obtain a restored image, ensuring that the resulting image remains visually coherent and unmarred by the removal process. ## 2 Request ### 1.1 HTTP Protocol HTTPS ### 1.2 Method Only support POST ### 1.3 charset utf-8 ### 1.4 Content-Type Content-Type: multipart/form-data ### 1.5 parameters

Parameters	Type	Required	Description	Value
img_file	file	No	Image File	1, Supported image types: jpg, jpeg, bmp, png 2, img_url Protocol: HTTP/HTTPS Max size: Up to 500 strings Timeout: Download Timeout of 20 seconds 3, One of three parameters must be used, priority: (1)img_file; (2)img_url; (3)img_b64; 4, Max Image Size The request image size should not exceed 10MB.
img_b64	string	No	Image Encoding in Base64 Format
img_url	string	No	Image download addres url, must use urlencode
img_format	int	No	Request image format	0: jpg/jpeg (default); 1: png; 2: bmp; When img_b64 or img_url is used, make sure img_format is correct
target_boxes	string	No	Target boxes: object will be removed or kept only in target boxes.	If there are many objects in the image, but you only want to remove/keep one or a few of them, you can set one or more target boxes, and only the objects in the target boxes will be removed/kept. The format of target boxes: x11,y11,x12,y12; x21,y21,x22,y22; Among them, (x11,y11,x12,y12) is box1 points, (x11,y11) is left-top point, (x12,y12) is right-bottom point; (x21,y21,x22,y22) is box2 points; You can set many boxes.
out_type	int	No	Option: returned image should use image url or base64 string	0: url(default); 1: base64 string
out_format	int	No	Option: returned image format; Mask image always use jpg format.	0: jpg(default); 1: png; When png is used, removed area of the image always is transparent
with_mask	int	No	Option: return mask image or not	0: don't return mask image(default); 1: return mask image If mask image is returned, it always use jpg format and the removed area is black
fill_color	int	No	Option: when out_format is jpg, the color is used to fill removed area	0: white color (default); 1: black color;
prompt	string	No	The prompt text is used to describe anything object that will be removed from an image	for example: skirt,hat Only used to API: remove-anything, remove-anything-impaint, keep-anything.

### 1.6 curl example ``` curl -X POST -F "img_file=@image.jpg" -F "out_type=0" -F "with_mask=1" -F "out_format=1" -F "fill_color=1" https://image-remove-and-impaint1.p.rapidapi.comapi/v1/img/remove-person ``` ## 2 Response ### 2.1 content-type Content-Type: application/json ### 2.2 fields

Parameters	Type	Required	Description	Value
img_url	string	No	result image url	if out_type is 0(url), img_url will be used
img_ext	string	Yes	result image format	if out_format is 0(jpg), img_ext is "jpg"; else img_ext is "png";
img_b64	string	No	result image base64 string	if out_type is 1(base64), img_b64 will be used
mask_url	string	No	result mask image url	if out_type is 0(url) and with_mask is 1(return mask), mask_url will be used
mask_ext	string	Yes	result mask image format	always is "jpg"
mask_b64	string	No	result image base64 string	if out_type is 1(base64) and with_mask is 1(return mask), mask_b64 will be used

### 2.3 example ``` { "img_b64": null, "img_ext": "jpg", "img_url": "http://img-global.seefine.cc/image/xxx.jpg?somesecurityparametes", "mask_b64": null, "mask_ext": "jpg", "mask_url": "http://img-global.seefine.cc/image/xxxmaks.jpg?somesecurityparametes" } ``` ## 3 http response code

http response code	describe
200	success
401	token is bad
403	request parameter is bad
500	system error