DOCUMENTATION
Chooch Computer Vision API Guide
Computer Vision API
Now you can accelerate your computer vision deployments by taking advantage of the Chooch computer vision API. Each imaging system captures its own images or videos, and then sends them to a hosted inference engine running AI models. Using Chooch’s computer vision API, these two systems can establish a standard protocol for communication. The host machine then takes the input image or video, runs it through the model, and sends the results back to the requester within a fraction of a second.
API v2
Object Detection API
URL POST
- Python
- Javascript (REST)
- cURL
import requests
import base64
import io
from PIL import Image
import json
image = Image.open("your_local_image").convert("RGB")
# Convert to a JPEG Buffer
buffered = io.BytesIO()
image.save(buffered, quality=90, format="JPEG")
# Base 64 Encode
base64_string = base64.b64encode(buffered.getvalue())
base64_string = base64_string.decode("ascii")
payload = { "base64str": base64_string, "model_id": '7f0829c0-fe45-4417-a8a0-1fb2c0b62d6d', "conf_thresh": 0.4, "nms_thresh": 0.45 }
api_url = 'https://apiv2.chooch.ai/predict?api_key=dd325728-54c0-45fc-82ad-b4ed037adcfd'
response = requests.put(api_url, data=json.dumps(payload))
json_data = json.loads(response.content)
print(json_data)
const axios = require("axios");
const fs = require("fs");
const image_str = fs.readFileSync("your_local_image", { encoding: "base64" });
const payload_json = JSON.stringify({ base64str: image_str, model_id: '7f0829c0-fe45-4417-a8a0-1fb2c0b62d6d', conf_thresh: 0.4, nms_thresh: 0.45 });
axios({ method: "PUT", url: "https://apiv2.chooch.ai/predict?api_key=dd325728-54c0-45fc-82ad-b4ed037adcfd", data: payload_json, headers: { "Content-Type": "application/json" } })
.then(function(response) { console.log(response.data); })
.catch(function(error) { console.log(error.message); });
base64_string = $(base64 "your_local_image")
curl -d '{"base64str": $base64_string, "model_id": "7f0829c0-fe45-4417-a8a0-1fb2c0b62d6d", "conf_thresh": 0.4, "nms_thresh": 0.45}' -H "Content-Type: application/json" -X PUT https://apiv2.chooch.ai/predict?api_key=dd325728-54c0-45fc-82ad-b4ed037adcfd
The response fields are:
model_id: The UID of the model. In this example it is 027765184-5b08-410c-97c7-c2caf3p01ad
apikey: API key provided by Chooch. You can get yours by signing in to Chooch AI Vision Studio. In this example it is 78a8f601-74e5-351b-11d6-6d80362a38b7
conf_thresh: Threshold for minimum confidence the model has on a detected object(box confidence score).
nms_thresh: Non-Maximum Suppression (NMS) threshold to select the best bounding box for an object and reject or “suppress” all other bounding boxes.
API v1
General Recognition AI API
The Chooch API is organized around REST. The API recognizes objects and concepts in videos and images from our pre-trained models.
The API is compatible with livestreams and live tagging.
Image Recognition API
The image recognition API is a straightforward REST API conducting classification and object detection predictions on images. The image URL is passed as a field to the API and the API returns a JSON based response with relevant predictions.
LOCAL IMAGE POST
- Python
- jQuery
- PHP
- cURL
# Image Recognition with Local Image
import requests
import json
url = 'https://api.chooch.ai/predict/image?apikey=346g5717-1sd3-35h6-9104-b8h5c819dn19'
files = {'image': open('local_image.jpg', 'rb')}
response = requests.post(url, files=files)
print(response.content)
<?php
$curl = curl_init();
curl_setopt_array($curl, array(
CURLOPT_URL => "https://api.chooch.ai/predict/image?apikey=346g5717-1sd3-35h6-9104-b8h5c819dn19",
CURLOPT_RETURNTRANSFER => true,
CURLOPT_ENCODING => "",
CURLOPT_MAXREDIRS => 10,
CURLOPT_TIMEOUT => 0,
CURLOPT_FOLLOWLOCATION => false,
CURLOPT_HTTP_VERSION => CURL_HTTP_VERSION_1_1,
CURLOPT_CUSTOMREQUEST => "POST",
CURLOPT_POSTFIELDS => array('image'=> new CURLFILE('')),
));
$response = curl_exec($curl);
$err = curl_error($curl);
curl_close($curl);
if ($err) {
echo "cURL Error #:" . $err;
} else {
echo $response;
} ?>
curl --location --request POST "https://api.chooch.ai/predict/image?apikey=346g5717-1sd3-35h6-9104-b8h5c819dn19" --form "image=@local_image.jpg"
URL POST
- Python
- jQuery
- PHP
- cURL
# Image Recognition with Image Url
import requests
import json
url = 'https://api.chooch.ai/predict/image?url=https://s3.amazonaws.com/choochdashboard/base_site/ronaldo_suit.jpg&apikey=346g5717-1sd3-35h6-9104-b8h5c819dn19'
response = requests.post(url)
print(response.content)
curl --location --request GET "https://api.chooch.ai/predict/image?url=https://s3.amazonaws.com/choochdashboard/base_site/ronaldo_suit.jpg&apikey=346g5717-1sd3-35h6-9104-b8h5c819dn19" --header "Content-Type: application/json"
API input fields are:
url: URL of the image file. In this example it is https://s3.amazonaws.com/choochdashboard/base_site/ronaldo_suit.jpg
apikey: API key provided by Chooch. You can get yours by signing in to Chooch AI Vision Studio. In this example it is 346g5717-1sd3-35h6-9104-b8h5c819dn19
SAMPLE OUTPUT (IMAGE)
The sample JSON response is:
Response fields are:
url: URL of the image file posted
status: Status of the API request. When successful the status is “ok”. For error messages see last section.
predictions: Predictions made on the image. Predictions are provided as a pst and have class_title and order fields.class_title is the name of the class predicted, and order is the order of relevancy of the particular class.
sub_predictions: Predictions made under dense classification. The image or frame is segmented into parts and classified based on the segments.
text_value: Text predictions.
face_name: Facial recognition name predictions.
face_count: Sum of faces.
coordinates: Coordinates of the object based on pixels. The format for the coordinates is X1, X2, Y1, Y2.
count: The number of times an object or concept appears in an image or a frame.
fileid: Unique File ID created for the image. The File ID is to identify the image for future references.
In this example it is ec7995bb557d4416ab97b819917fc18c
GENERAL STATUS MESSAGES
ok
Error (Wrong Parameters)
Error (Invapd API Key)
Error (You Have Reached Your Trial API Call pmit)
Error (No File)
FTP DROP FUNCTION
The FTP drop function was developed on top of the Chooch API for image and video professionals who want to tag their images and videos autonomously. For every cpent a personal FTP host is issued together with the API.
FTP Host: ftp://52.207.237.100
API Key: 6werc42-1675-6553-8105
The FTP drop sends the image to the API. The API predicts the image and places the results in the keywords field of the IPTC metadata file. Once the metadata has been successfully placed into the keywords fields of the IPTC file, it is sent to the destination directed by the user.
Custom Image Recognition API
LOCAL IMAGE POST
- Python
- jQuery
- PHP
- cURL
# Image Recognition with Local Image
import requests
import json
url = 'https://api.chooch.ai/predict/image?apikey=346g5717-1sd3-35h6-9104-b8h5c819dn19&model_id=808'
files = {'image': open('local_image.jpg', 'rb')}
response = requests.post(url, files=files)
print(response.content)
<?php
$curl = curl_init();
curl_setopt_array($curl, array(
CURLOPT_URL => "https://api.chooch.ai/predict/image?apikey=346g5717-1sd3-35h6-9104-b8h5c819dn19&model_id=808",
CURLOPT_RETURNTRANSFER => true,
CURLOPT_ENCODING => "",
CURLOPT_MAXREDIRS => 10,
CURLOPT_TIMEOUT => 0,
CURLOPT_FOLLOWLOCATION => false,
CURLOPT_HTTP_VERSION => CURL_HTTP_VERSION_1_1,
CURLOPT_CUSTOMREQUEST => "POST",
CURLOPT_POSTFIELDS => array('image'=> new CURLFILE('')),
));
$response = curl_exec($curl);
$err = curl_error($curl);
curl_close($curl);
if ($err) {
echo "cURL Error #:" . $err;
} else {
echo $response;
} ?>
curl --location --request POST "https://api.chooch.ai/predict/image?apikey=346g5717-1sd3-35h6-9104-b8h5c819dn19&model_id=808" --form "image=@local_image.jpg"
URL POST
- Python
- jQuery
- PHP
- cURL
# Image Recognition with Image Url
import requests
import json
url = 'https://api.chooch.ai/predict/image?url=https://chooch-share.s3.amazonaws.com/cat.jpg&apikey=346g5717-1sd3-35h6-9104-b8h5c819dn19&model_id=808'
response = requests.post(url)
print(response.content)
curl --location --request GET "https://api.chooch.ai/predict/image?url=https://chooch-share.s3.amazonaws.com/cat.jpg&apikey=346g5717-1sd3-35h6-9104-b8h5c819dn19&model_id=808" --header "Content-Type: application/json"
API input fields are:
model_id: The id of the model you trained.In this example it is 808
url: URL of the image file. In this example it is https://chooch-share.s3.amazonaws.com/cat.jpg
apikey: API key provided by Chooch. You can get yours by signing in to Chooch AI Vision Studio. In this example it is 346g5717-1sd3-35h6-9104-b8h5c819dn19
SAMPLE OUTPUT (IMAGE)
The sample JSON response is:
{
"status": "ok",
"model_id": "808",
"url": "https://chooch-share.s3.amazonaws.com/cat.jpg",
"predictions": [
{
"class_title": "cat",
"order": 1
}
],
"prediction_type": "image"
}
The response fields are:
status: Status of the API request.
model_id: The id of the model you trained.In this example it is 808
url: URL of the image file posted. In this example it is https://chooch-share.s3.amazonaws.com/cat.jpg
predictions: Predictions made on the image. Predictions are provided as a pst and have class_title and order fields.class_title is the name of the class predicted, and order is the order of relevancy of the particular class.
predict_type: Predict type is image.
Custom Object Detection API
URL POST
- Python
- jQuery
- PHP
- cURL
# Object Detection with Image Url
import requests
import json
url = 'https://api.chooch.ai/predict/object_detection/?url=https://choochdashboard.s3.amazonaws.com/truck.jpg&apikey=346g5717-1sd3-35h6-9104-b8h5c819dn19c&model_id=777'
response = requests.post(url)
json_data = json.loads(response.content)
print(json_data)
curl --location --request GET "https://api.chooch.ai/predict/object_detection/?url=https://choochdashboard.s3.amazonaws.com/truck.jpg&apikey=346g5717-1sd3-35h6-9104-b8h5c819dn19c&model_id=777"
LOCAL IMAGE POST
- Python
- jQuery
- PHP
- cURL
import requests
import json
import time
url = 'https://api.chooch.ai/predict/object_detection/?model_id=777&apikey=346g5717-1sd3-35h6-9104-b8h5c819dn19'
files = {'image': open('local_image.jpg', 'rb')}
response = requests.post(url, files=files)
json_data = json.loads(response.content)
print(json_data)
curl --location --request POST "https://api.chooch.ai/predict/object_detection/?model_id=777&apikey=346g5717-1sd3-35h6-9104-b8h5c819dn19" --form "image=@local_image.jpg"
SAMPLE RESPONSE:
{
"status": "ok",
"prediction_type": "object_detection",
"predictions": [
{
"class_title": "person",
"model_id": 777,
"score": 0.74613,
"coordinates": {
"xmin": 0,
"ymin": 35,
"ymax": 297,
"xmax": 179
}
}
]
}
API input fields are:
model_id: The id of the model you trained.In this example it is 777
coordinates: Coordinates of the object based on pixels. The format for the coordinates is xmin, ymin ,xmax, ymax
Custom Facial Recognition API
Chooch Face Recognition consists of Perception > People > Images
You can search and tag a face in the Chooch API by feeding the Chooch API a Face Image. Below is a sample post of an image url.
SAMPLE OUTPUT (IMAGE)
The fields are:
url: This is the image url of the face to be searched.In this example it is https://choochdashboard.s3.amazonaws.com/base_site/1461123713-esq050116cover001s.jpg
person_id_filter: This field is a filter that is optional to pass. The default -1 which means don’t filter by person. If a valid person_id is passed, Chooch will search only in that given person’s data.
In this example it is -1
>model_id: This is the id of the perception that the search will be made on. This field is required.
In this example it is 14
You can also post an image through the image field. Below are 2 separate sample posts in python.
URL POST
- Python
- jQuery
- PHP
- cURL
import requests
import json
import time
url = 'https://api.chooch.ai/predict/face?url=https://choochdashboard.s3.amazonaws.com/base_site/1461123713-esq050116cover001s.jpg&person_id_filter=-1&model_id=14&apikey=346g5717-1sd3-35h6-9104-b8h5c819dn19'
response = requests.post(url)
json_data = json.loads(response.content)
print(json_data)
curl --location --request POST "https://api.chooch.ai/predict/face?url=https://choochdashboard.s3.amazonaws.com/base_site/1461123713-esq050116cover001s.jpg&person_id_filter=-1&model_id=14&apikey=346g5717-1sd3-35h6-9104-b8h5c819dn19"
LOCAL IMAGE POST
- Python
- jQuery
- PHP
- cURL
- Java
- C#
import requests
import json
import time
url = 'https://api.chooch.ai/predict/face?person_id_filter=-1&model_id=5&apikey=346g5717-1sd3-35h6-9104-b8h5c819dn19'
files = {'image': open('local_image.jpg', 'rb')}
response = requests.post(url, files=files)
json_data = json.loads(response.content)
print(json_data)
curl --location --request POST "https://api.chooch.ai/predict/face?person_id_filter=-1&model_id=14&apikey=346g5717-1sd3-35h6-9104-b8h5c819dn19" --form "image=@local_image.jpg"
JSON RESPONSE
The example JSON response is here below. In this example, the similarity is 0.80. The face_recog_hit is True when there is a match and False when there is no match.
This API function allows the user to add a face image with a unique id (it can be a unique person name, or a unique id assigned to a person):
- Python
- jQuery
- PHP
- cURL
- Java
- C#
import requests
import json
import time
url = 'https://api.chooch.ai/predict/face?&key_id=3245&model_id=7&command=insert_person_image_key_id&apikey=46g5717-1sd3-35h6-9104-b8h5c819dn19'
files = {'image': open('/home/ashwmadhu/Desktop/12.jpg', 'rb')}
response = requests.post(url, files=files)
json_data = json.loads(response.content)
print(json_data)
curl -X POST 'https://api.chooch.ai/predict/face?key_id=3245&model_id=7&command=insert_person_image_key_id&apikey=46g5717-1sd3-35h6-9104-b8h5c819dn19' -F image=@/home/ashwmadhu/Desktop/12.jpg
JSON RESPONSE
{
"status_description":"success",
"post_type":"insert_person_image",
"status":41868
}
CREATE PERSON API COMMAND
You can create a person through the Chooch API. Below is a sample Python based usage of the create_person command.
- Python
- jQuery
- PHP
- cURL
- Java
- C#
# Create Person Python Code
import requests
import json
url = 'https://api.chooch.ai/predict/face?person_name=Tom Bill&model_id=5&apikey=cx065e5d-51wc87-9df6-163e-8b65a78k51e3&command=create_person'
response = requests.post(url)
json_data = json.loads(response.content)
print(json_data)
curl --location --request POST "https://api.chooch.ai/predict/face?person_name=Tom%20Bill&model_id=5&apikey=cx065e5d-51wc87-9df6-163e-8b65a78k51e3&command=create_person"
JSON RESPONSE
{
person_id: 37,
status: 37,
status_description: success,
post_type: create_person
}
RETURN STATUS VALUES:
status > 0 : success, and the return value is the person_id.
status = 0 : error has occurred.
status = -1 : invalid model id.
status = -2 : invalid person name.
p
INSERT PERSON IMAGE API COMMAND
A face image can be added to an existing person using the person_id. Below is a sample Python based usage of the insert_person_image command.
- Python
- jQuery
- PHP
- cURL
# Add face image to person with person_id_filter (person_id) Python Code
import requests
import json
url = 'https://api.chooch.ai/predict/face?person_id_filter=37&apikey=cx065e5d-51wc87-9df6-163e-8b65a78k51e3&command=insert_person_image'
files = {'image': open('your_image.jpg', 'rb')}
response = requests.post(url, files=files)
json_data = json.loads(response.content)
print(json_data)
curl --location --request POST "https://api.chooch.ai/predict/face?person_id_filter=37&apikey=cx065e5d-51wc87-9df6-163e-8b65a78k51e3&command=insert_person_image" --form "image=@"
JSON RESPONSE
{
status: 210,
status_description: success,
post_type: insert_person_image
}
RETURN STATUS VALUES:
status > 0 : success, and the return value is the image id
status = 0 : error has occurred.
status = -1 : no face detected.
status = -2 : invalid person id.
Thank you for evaluating the Chooch AI computer vision API. If you don’t already have a Chooch AI Vision Studio account, we encourage you to create a free account.