app.py

import gradio as gr
import torch
import torch.nn.functional as F
import numpy as np
from transformers import SegformerImageProcessor, SegformerForSemanticSegmentation
from PIL import Image
import os 
from functools import partial


def resize_image(image, target_size=1024):
    h_img, w_img = image.size
    if h_img < w_img:
        new_h, new_w = target_size, int(w_img * (target_size / h_img))
    else:
        new_h, new_w  = int(h_img * (target_size / w_img)), target_size
    
    resized_img = image.resize((new_h, new_w))
    return resized_img

def segment_image(image, preprocessor, model, crop_size = (1024, 1024), num_classes = 40):
    print(type(image))
    
    h_crop, w_crop = crop_size
    print(image.size)
    
    img = torch.Tensor(np.array(resize_image(image, target_size=1024)).transpose(2, 0, 1)).unsqueeze(0).to(device)
    batch_size, _, h_img, w_img = img.size()
    print(img.size())
    
    h_grids = int(np.round(3/2*h_img/h_crop)) if h_img > h_crop else 1
    w_grids = int(np.round(3/2*w_img/w_crop)) if w_img > w_crop else 1
    print(h_grids, w_grids)
    
    h_stride = int((h_img - h_crop + h_grids -1)/(h_grids -1)) if h_grids > 1 else h_crop
    w_stride = int((w_img - w_crop + w_grids -1)/(w_grids -1)) if w_grids > 1 else w_crop
    print(h_stride, w_stride)
    
    preds = img.new_zeros((batch_size, num_classes, h_img, w_img))
    count_mat = img.new_zeros((batch_size, 1, h_img, w_img))
    
    for h_idx in range(h_grids):
        for w_idx in range(w_grids):
          y1 = h_idx * h_stride
          x1 = w_idx * w_stride
          y2 = min(y1 + h_crop, h_img)
          x2 = min(x1 + w_crop, w_img)
          y1 = max(y2 - h_crop, 0)
          x1 = max(x2 - w_crop, 0)
          crop_img = img[:, :, y1:y2, x1:x2]
          print(x1, x2, y1, y2)
          with torch.no_grad():
            inputs = preprocessor(crop_img, return_tensors = "pt")
            outputs = model(**inputs)
        
          resized_logits = F.interpolate(
              outputs.logits[0].unsqueeze(dim=0), size=crop_img.shape[-2:], mode="bilinear", align_corners=False
          )
          preds += F.pad(resized_logits,
                          (int(x1), int(preds.shape[3] - x2), int(y1),
                          int(preds.shape[2] - y2)))
          count_mat[:, :, y1:y2, x1:x2] += 1
    
    assert (count_mat == 0).sum() == 0
    preds = preds / count_mat
    
    preds = preds.argmax(dim=1)
    
    preds = F.interpolate(preds.unsqueeze(0).type(torch.uint8), size=image.size[::-1], mode='nearest')
    label_pred = preds.squeeze().cpu().numpy()
    
    # label_pred_colors =  np.array([[id2color[pixel] for pixel in row] for row in np.array(label_pred)])
    # mask_image = Image.fromarray(label_pred_colors.astype(np.uint8), 'RGB')
    # overlay = Image.blend(image.convert("RGBA"), mask_image.convert("RGBA"), alpha=0.6)
    # return overlay

    seg_info = [(label_pred == int(id), label) for id, label in id2label.items()]
    return (image, seg_info)



# # Create Gradio interface
# interface = gr.Interface(
#     fn=segment_image,
#     inputs=[gr.Image(type="pil")],
#     outputs=[gr.Image(type="pil")],
#     title="Coral Segmentation with SegFormer",
#     description="Official demo for **Coralscapes**",
#     examples=example_files 
# )

# # Launch the demo
# interface.launch()


if __name__ == "__main__":
    device = 'cuda' if torch.cuda.is_available() else 'cpu'
    print(device)
    
    # Load model and processor
    preprocessor = SegformerImageProcessor.from_pretrained("EPFL-ECEO/segformer-b2-finetuned-coralscapes-1024-1024")
    model = SegformerForSemanticSegmentation.from_pretrained("EPFL-ECEO/segformer-b2-finetuned-coralscapes-1024-1024").to(device)
    model.eval()
 
    id2label = {"1": "seagrass", "2": "trash", "3": "other coral dead", "4": "other coral bleached", "5": "sand", "6": "other coral alive", "7": "human", "8": "transect tools", "9": "fish", "10": "algae covered substrate", "11": "other animal", "12": "unknown hard substrate", "13": "background", "14": "dark", "15": "transect line", "16": "massive/meandering bleached", "17": "massive/meandering alive", "18": "rubble", "19": "branching bleached", "20": "branching dead", "21": "millepora", "22": "branching alive", "23": "massive/meandering dead", "24": "clam", "25": "acropora alive", "26": "sea cucumber", "27": "turbinaria", "28": "table acropora alive", "29": "sponge", "30": "anemone", "31": "pocillopora alive", "32": "table acropora dead", "33": "meandering bleached", "34": "stylophora alive", "35": "sea urchin", "36": "meandering alive", "37": "meandering dead", "38": "crown of thorn", "39": "dead clam"}
    label2color = {"human": [255, 0, 0], "background": [29, 162, 216], "fish": [255, 255, 0], "sand": [194, 178, 128], "rubble": [161, 153, 128], "unknown hard substrate": [125, 125, 125], "algae covered substrate": [125, 163, 125], "dark": [31, 31, 31], "branching bleached": [252, 231, 240], "branching dead": [123, 50, 86], "branching alive": [226, 91, 157], "stylophora alive": [255, 111, 194], "pocillopora alive": [255, 146, 150], "acropora alive": [236, 128, 255], "table acropora alive": [189, 119, 255], "table acropora dead": [85, 53, 116], "millepora": [244, 150, 115], "turbinaria": [228, 255, 119], "other coral bleached": [250, 224, 225], "other coral dead": [114, 60, 61], "other coral alive": [224, 118, 119], "massive/meandering alive": [236, 150, 21], "massive/meandering dead": [134, 86, 18], "massive/meandering bleached": [255, 248, 228], "meandering alive": [230, 193, 0], "meandering dead": [119, 100, 14], "meandering bleached": [251, 243, 216], "transect line": [0, 255, 0], "transect tools": [8, 205, 12], "sea urchin": [0, 142, 255], "sea cucumber": [0, 231, 255], "anemone": [0, 255, 189], "sponge": [240, 80, 80], "clam": [189, 255, 234], "other animal": [0, 255, 255], "trash": [255, 0, 134], "seagrass": [125, 222, 125], "crown of thorn": [179, 245, 234], "dead clam": [89, 155, 134]}
    label2colorhex = {k:'#%02x%02x%02x' % tuple(v) for k,v in label2color.items()}
    print(label2colorhex)

    with gr.Blocks(title="Coral Segmentation with SegFormer") as demo:
        gr.Markdown("""<h1><center>Coral Segmentation with SegFormer</center></h1>""") 
        with gr.Row():
            img_input = gr.Image(type="pil", label="Input image")
            # img_output = gr.Image(type="pil", label="Predictions")
            img_output = gr.AnnotatedImage(label="Predictions", color_map=label2colorhex)

        section_btn = gr.Button("Segment Image")
        section_btn.click(partial(segment_image, preprocessor=preprocessor, model=model), img_input, img_output)
        
        example_files = os.listdir('assets/examples')
        example_files.sort()
        print(example_files)
        example_files = [os.path.join('assets/examples', filename) for filename in example_files]

        gr.Examples(examples=example_files, inputs=img_input, outputs=img_output)

    demo.launch()