1

My project aims to detect object labels and coordinates and then convert them into a string which is converted into voice using gTTS but I keep getting an attribute error in the prediction labels. I am new to this framework, any help will be appreciated.

Code:

import cv2
from gtts import gTTS
import os
from ultralytics import YOLO

def convert_labels_to_text(labels):
    text = ", ".join(labels)
    return text

class YOLOWithLabels(YOLO):
    def __call__(self, frame):
        results = super().__call__(frame)
        labels = results.pred[0].get_field("labels").tolist()
        annotated_frame = results.render()
        return annotated_frame, labels

cap = cv2.VideoCapture(0)
model = YOLOWithLabels('yolov8n.pt')

while cap.isOpened():
    success, frame = cap.read()

    if success:
        annotated_frame, labels = model(frame)

        message = convert_labels_to_text(labels)

        tts_engine = gTTS(text=message)  # Initialize gTTS with the message

        tts_engine.save("output.mp3")
        os.system("output.mp3")

        cv2.putText(annotated_frame, message, (10, 30), cv2.FONT_HERSHEY_SIMPLEX, 1, (0, 255, 0), 2)
        cv2.imshow("YOLOv8 Inference", annotated_frame)

        if cv2.waitKey(1) & 0xFF == ord("q"):
            break

    else:
        break

cap.release()
cv2.destroyAllWindows()

Error

File "C:\Users\alien\Desktop\YOLOv8 project files\gtts service\testservice.py", line 13, in __call__
    labels = results.pred[0].get_field("labels").tolist()
             ^^^^^^^^^^^^
AttributeError: 'list' object has no attribute 'pred'

print(results)

orig_shape: (480, 640)
path: 'image0.jpg'
probs: None
save_dir: None
speed: {'preprocess': 3.1604766845703125, 'inference': 307.905912399292, 'postprocess': 2.8924942016601562}]
0: 480x640 1 person, 272.4ms
Speed: 3.0ms preprocess, 272.4ms inference, 4.0ms postprocess per image at shape (1, 3, 640, 640)
[ultralytics.yolo.engine.results.Results object with attributes:
    boxes: ultralytics.yolo.engine.results.Boxes object
    keypoints: None
    keys: ['boxes']
    masks: None
    names: {0: 'person', 1: 'bicycle', 2: 'car', 3: 'motorcycle', 4: 'airplane', 5: 'bus', 6: 'train', 7: 'truck', 8: 'boat', 9: 'traffic light', 10: 'fire hydrant', 11: 'stop sign', 12: 'parking meter', 13: 'bench', 14: 'bird', 15: 'cat', 16: 'dog', 17: 'horse', 18: 'sheep', 19: 'cow', 20: 'elephant', 21: 'bear', 22: 'zebra', 23: 'giraffe', 24: 'backpack', 25: 'umbrella', 26: 'handbag', 27: 'tie', 28: 'suitcase', 29: 'frisbee', 30: 'skis', 31: 'snowboard', 32: 'sports ball', 33: 'kite', 34: 'baseball bat', 35: 'baseball glove', 36: 'skateboard', 37: 'surfboard', 38: 'tennis racket', 39: 'bottle', 40: 'wine glass', 41: 'cup', 42: 'fork', 43: 'knife', 44: 'spoon', 45: 'bowl', 46: 'banana', 47: 'apple', 48: 'sandwich', 49: 'orange', 50: 'broccoli', 51: 'carrot', 52: 'hot dog', 53: 'pizza', 54: 'donut', 55: 'cake', 56: 'chair', 57: 'couch', 58: 'potted plant', 59: 'bed', 60: 'dining table', 61: 'toilet', 62: 'tv', 63: 'laptop', 64: 'mouse', 65: 'remote', 66: 'keyboard', 67: 'cell phone', 68: 'microwave', 69: 'oven', 70: 'toaster', 71: 'sink', 72: 'refrigerator', 73: 'book', 74: 'clock', 75: 'vase', 76: 'scissors', 77: 'teddy bear', 78: 'hair drier', 79: 'toothbrush'}
    orig_img: array([[[168, 167, 166],
            [165, 165, 165],
            [165, 166, 167],
            ...,
            [183, 186, 178],
            [183, 186, 178],
            [184, 187, 179]],

           [[168, 167, 165],
            [166, 165, 165],
            [166, 167, 166],
            ...,
            [184, 187, 179],
            [183, 186, 178],
            [184, 187, 179]],

           [[168, 167, 164],
            [167, 167, 164],
            [167, 167, 165],
            ...,
            [184, 187, 178],
            [184, 187, 179],
            [183, 186, 178]],

           ...,

           [[196, 192, 185],
            [196, 192, 185],
            [196, 192, 185],
            ...,
            [ 25,  29,  38],
            [ 22,  25,  35],
            [ 20,  24,  34]],

           [[199, 195, 187],
            [197, 193, 186],
            [197, 193, 186],
            ...,
            [ 23,  26,  35],
            [ 22,  25,  35],
            [ 22,  25,  35]],

           [[199, 195, 187],
            [199, 195, 187],
            [199, 195, 187],
            ...,
            [ 20,  24,  33],
            [ 19,  23,  33],
            [ 19,  23,  33]]], dtype=uint8)
greybeard
  • 2,249
  • 8
  • 30
  • 66
asfriendlyascarbon
  • 13
  • 4
  • Could you try: `labels = results[0].get_field("labels").tolist()` – doneforaiur Jun 21 '23 at 17:38
  • Thank you for your suggestion kind sir, but it seems to throw off another attribute based error: AttributeError: 'Results' object has no attribute 'get_field'. – asfriendlyascarbon Jun 21 '23 at 17:41
  • I'm not familiar with either YOLO nor Google TTS, but; https://github.com/orgs/ultralytics/discussions/2547#discussioncomment-5889458 Could you try this example? I'm not sure how declaring a ``YOLOWithLabels` class would help. Also, couldn't find anything about `labels` here; https://docs.ultralytics.com/reference/yolo/engine/results/#ultralytics.yolo.engine.results.Probs.top1conf But will keep looking. – doneforaiur Jun 21 '23 at 17:59
  • Ooo, found something interesting. `YOLO('yolo8n-cls.pt')` gives a model that only classifies what is visible in the image. https://docs.ultralytics.com/tasks/classify/ . In the meantime, can you just `print(results)` and post the output? I'm really confused as their documentation is really weird. – doneforaiur Jun 21 '23 at 18:13
  • Thank you so much! I will go through the documentation even though it feels difficult to understand. I have attached print(results) output to the post) – asfriendlyascarbon Jun 22 '23 at 05:22
  • Okay, we can clearly see it has detected `1 person`. You are very close to the solution, sadly I have to go but I'm sure you'll get an answer in no time! ^^ – doneforaiur Jun 22 '23 at 05:43
  • 1
    Yes, but I'm quite unsure of how to extract that from all the other data and the arrays which I don't need. I get that you have to go :( Thanks for the help so far. – asfriendlyascarbon Jun 22 '23 at 05:47
  • Try; `for result in results: print(result)`. That might also give you a clue. – doneforaiur Jun 22 '23 at 05:56
  • I'm still stuck and it seems you're my last hope :3 – asfriendlyascarbon Jun 22 '23 at 13:24

1 Answers1

0

May God have mercy on whoever wrote Ultralytics's docs... Here's how you can print only the labels:

from ultralytics import YOLO

model = YOLO('yolov8n.pt')
results = model('http://images.cocodataset.org/val2017/000000397133.jpg')

print(model.names)

for result in results:
    boxes = result.boxes.cpu().numpy()
    for box in boxes:
        print(model.names[box.cls[0]])

model.names contains all the classes that can be predicted. Each box has a cls (class for short) attribute which is a list of int values. You can search for that class in the model.names dictionary.

PS: Each box has a list of ints which indicates that the model can return multiple classes for one bounding box. This example only takes first of the classes in the box.cls list.

doneforaiur
  • 1,308
  • 7
  • 14
  • 21