File size: 3,145 Bytes
2fe64fd
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
from typing import Tuple, List

import cv2
import numpy as np
import supervision as sv
import torch
from PIL import Image
from torchvision.ops import box_convert

import groundingdino.datasets.transforms as T
from groundingdino.models import build_model
from groundingdino.util.misc import clean_state_dict
from groundingdino.util.slconfig import SLConfig
from groundingdino.util.utils import get_phrases_from_posmap


def preprocess_caption(caption: str) -> str:
    result = caption.lower().strip()
    if result.endswith("."):
        return result
    return result + "."


def load_model(model_config_path: str, model_checkpoint_path: str, device: str = "cuda"):
    args = SLConfig.fromfile(model_config_path)
    args.device = device
    model = build_model(args)
    checkpoint = torch.load(model_checkpoint_path, map_location="cpu")
    model.load_state_dict(clean_state_dict(checkpoint["model"]), strict=False)
    model.eval()
    return model


def load_image(image_path: str) -> Tuple[np.array, torch.Tensor]:
    transform = T.Compose(
        [
            T.RandomResize([800], max_size=1333),
            T.ToTensor(),
            T.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225]),
        ]
    )
    image_source = Image.open(image_path).convert("RGB")
    image = np.asarray(image_source)
    image_transformed, _ = transform(image_source, None)
    return image, image_transformed


def predict(
        model,
        image: torch.Tensor,
        caption: str,
        box_threshold: float,
        text_threshold: float,
        device: str = "cuda"
) -> Tuple[torch.Tensor, torch.Tensor, List[str]]:
    caption = preprocess_caption(caption=caption)

    model = model.to(device)
    image = image.to(device)

    with torch.no_grad():
        outputs = model(image[None], captions=[caption])

    prediction_logits = outputs["pred_logits"].cpu().sigmoid()[0]  # prediction_logits.shape = (nq, 256)
    prediction_boxes = outputs["pred_boxes"].cpu()[0]  # prediction_boxes.shape = (nq, 4)

    mask = prediction_logits.max(dim=1)[0] > box_threshold
    logits = prediction_logits[mask]  # logits.shape = (n, 256)
    boxes = prediction_boxes[mask]  # boxes.shape = (n, 4)

    tokenizer = model.tokenizer
    tokenized = tokenizer(caption)

    phrases = [
        get_phrases_from_posmap(logit > text_threshold, tokenized, tokenizer).replace('.', '')
        for logit
        in logits
    ]

    return boxes, logits.max(dim=1)[0], phrases


def annotate(image_source: np.ndarray, boxes: torch.Tensor, logits: torch.Tensor, phrases: List[str]) -> np.ndarray:
    h, w, _ = image_source.shape
    boxes = boxes * torch.Tensor([w, h, w, h])
    xyxy = box_convert(boxes=boxes, in_fmt="cxcywh", out_fmt="xyxy").numpy()
    detections = sv.Detections(xyxy=xyxy)

    labels = [
        f"{phrase} {logit:.2f}"
        for phrase, logit
        in zip(phrases, logits)
    ]

    box_annotator = sv.BoxAnnotator()
    annotated_frame = cv2.cvtColor(image_source, cv2.COLOR_RGB2BGR)
    annotated_frame = box_annotator.annotate(scene=annotated_frame, detections=detections, labels=labels)
    return annotated_frame