app.py

import cv2
import gradio as gr
from huggingface_hub import hf_hub_download

from vision.ssd.mobilenet_v2_ssd_lite import (
    create_mobilenetv2_ssd_lite,
    create_mobilenetv2_ssd_lite_predictor,
)

MODEL_REPO = "fa0311/oita-ken-strawberries-mobilenet"
MODEL_FILENAME = "20250129_053504/mb2-ssd-lite-Epoch-55-Loss-1.508891262114048.pth"
LABELS_FILENAME = "20250129_053504/labels.txt"

model_path = hf_hub_download(repo_id=MODEL_REPO, filename=MODEL_FILENAME)
label_path = hf_hub_download(repo_id=MODEL_REPO, filename=LABELS_FILENAME)

with open(label_path, "r") as f:
    class_names = [name.strip() for name in f.readlines()]

net = create_mobilenetv2_ssd_lite(len(class_names), is_test=True)
net.load(model_path)
predictor = create_mobilenetv2_ssd_lite_predictor(net, candidate_size=200)


def detect_objects(image, threshold):
    image = cv2.cvtColor(image, cv2.COLOR_RGB2BGR)
    boxes, labels, probs = predictor.predict(image, 10, threshold)

    for i in range(boxes.size(0)):
        box = list(map(int, boxes[i, :]))
        cv2.rectangle(image, (box[0], box[1]), (box[2], box[3]), (255, 255, 0), 4)
        label = f"{class_names[labels[i]]}: {probs[i]:.2f}"
        cv2.putText(
            image,
            label,
            (box[0] + 10, box[1] + 25),
            cv2.FONT_HERSHEY_SIMPLEX,
            0.8,
            (255, 0, 255),
            2,
        )

    return cv2.cvtColor(image, cv2.COLOR_BGR2RGB)


iface = gr.Interface(
    fn=detect_objects,
    inputs=[
        gr.Image(type="numpy"),
        gr.Slider(0.1, 1.0, value=0.7, label="Detection Threshold"),
    ],
    outputs=gr.Image(type="numpy"),
    title="SSD Object Detection - Strawberry quality classification",
    description="Upload an image of strawberries to detect objects using MobileNetV2-SSD-Lite.",
)

if __name__ == "__main__":
    iface.launch()