image size + boxes infos

app.py

@@ -10,6 +10,7 @@ import cv2
 from unilm.dit.object_detection.ditod import add_vit_config
 
 import torch
+import numpy as np
 
 from detectron2.config import CfgNode as CN
 from detectron2.config import get_cfg
@@ -51,8 +52,18 @@ def analyze_image(img):
     result_image = result.get_image()[:, :, ::-1]
 
     num_instances = len(output)
+    image_size = output._image_size
+    fields = list(output.get_fields().keys())
+    for field in fields:
+        if field == 'pred_boxes':
+            boxes = output.get_fields()[field]
+            boxes_numpy = boxes.tensor.cpu().numpy()
+            boxes_bytes = boxes_numpy.tobytes()
+            boxes_numpy_shape = str(boxes_numpy.shape)
+            boxes_numpy_dtype = str(boxes_numpy.dtype)
+            # boxes_recover = torch.from_numpy(np.frombuffer(boxes_bytes, dtype=boxes_numpy_dtype).reshape(boxes_numpy_shape))
     
-    return result_image, num_instances
+    return result_image, num_instances, image_size, boxes_bytes, boxes_numpy_shape, boxes_numpy_dtype
     
 title = "Interactive demo: Document Layout Analysis with DiT"
 description = "Demo for Microsoft's DiT, the Document Image Transformer for state-of-the-art document understanding tasks. This particular model is fine-tuned on PubLayNet, a large dataset for document layout analysis (read more at the links below). To use it, simply upload an image or use the example image below and click 'Submit'. Results will show up in a few seconds. If you want to make the output bigger, right-click on it and select 'Open image in new tab'."
@@ -62,7 +73,14 @@ css = ".output-image, .input-image, .image-preview {height: 600px !important}"
 
 iface = gr.Interface(fn=analyze_image, 
                      inputs=gr.inputs.Image(type="numpy", label="document image"), 
-                     outputs=[gr.outputs.Image(type="numpy", label="annotated document"), gr.outputs.Textbox(label="num instances")],
+                     outputs=[
+                         gr.outputs.Image(type="numpy", label="annotated document"), 
+                         gr.outputs.Textbox(label="num instances"),
+                         gr.outputs.Textbox(label="image size (h,w in pixels)"),
+                         gr.outputs.Textbox(label="boxes bytes"),
+                         gr.outputs.Textbox(label="boxes numpy shape"),
+                         gr.outputs.Textbox(label="boxes numpy dtype"),
+                         ],
                      title=title,
                      description=description,
                      examples=examples,