search

chingkhei-th / chicken-weight-using-GroundingDINO-and_SAM

Chicken weight estimation using Grounding DINO, SAM and LinearRegression
MIT License
0 stars 0 forks source link

i want to know this full report Linear Regression model and data folder #1

Open noorkhokhar99 opened 2 months ago

noorkhokhar99 commented 2 months ago

Must have a Linear Regression model (.pkl) file trained on chicken area-weight relationship to predict chicken weight.

how to we integration this part Is this accurate?

chingkhei-th commented 2 months ago

As I've mention, the linear regression model was trained by taking the area of the chicken (area getting by running instance segmentation on the chicken) as feature (input) and the weight of each chicken as target (output). This model will predict the weight of the chicken detected from the image.

To integrate this model you can modify the file path in the main.py 

# Load the linear regression model
linear_regression_model = joblib.load("./regression-model/LRmodel.pkl")

And for the accuracy, I would say it's not so accurate as there are many factors that affect the accuracy such as camera placement height when taking the data, image distortion by the camera, lack of dataset quantity, lack of proper hardware like (depth sense/ stereo camera to accurately detect the height and position of the chickens), etc. If you could play around and get rid of the factors that I've mention you could get more accurate result. But that would be hard and tedious task.