my bounding boxes are misplaced when inferencing with api but the roboflow visualize model shows correctly and the coordinates in the predictions returned by both are different .only if the predictions are same the bounding boxes will be correct when i calculate the edges .is this have to do something with the preprocessing steps applied . i have applied the autoorient and stretch to steps .wouldn’t they be applied when inferencing too
please resolve my issue
Hello, thank you for your message. Can you please provide an example of the discrepancy between the two outputs (API output and visualize tab output) as well as the image that results in the incorrect bounding box?
From the information you’ve described, I’m unfortunately not able to replicate the problem on my own test project, and would need more information to identify the best next step. Thank you.
hey thank you for reverting back this is the api predictions response {
“inference_id”: “491d98c8-76fb-49ea-9ba4-c58fffb14515”,
“time”: 0.04426114500074618,
“image”: {
“width”: 998,
“height”: 643
},
“predictions”: [
{
“x”: 364.5,
“y”: 131.5,
“width”: 69.0,
“height”: 57.0,
“confidence”: 0.8628576397895813,
“class”: “trees”,
“class_id”: 0,
“detection_id”: “219ee835-a3c1-470d-b07f-95b038bf72bf”
},
{
“x”: 359.0,
“y”: 221.0,
“width”: 92.0,
“height”: 98.0,
“confidence”: 0.8393381237983704,
“class”: “trees”,
“class_id”: 0,
“detection_id”: “38866e22-86ea-4f06-9460-dbf066799cb0”
},
{
“x”: 331.0,
“y”: 456.5,
“width”: 94.0,
“height”: 91.0,
“confidence”: 0.8388462662696838,
“class”: “trees”,
“class_id”: 0,
“detection_id”: “ffdca9b2-a723-4b84-ad7d-59c9f4024e21”
},
{
“x”: 703.0,
“y”: 356.0,
“width”: 186.0,
“height”: 222.0,
“confidence”: 0.8291531801223755,
“class”: “trees”,
“class_id”: 0,
“detection_id”: “ee51aee7-158d-470c-8ae9-86e7cf4d82a4”
},
{
“x”: 413.0,
“y”: 68.0,
“width”: 60.0,
“height”: 60.0,
“confidence”: 0.8286166191101074,
“class”: “trees”,
“class_id”: 0,
“detection_id”: “e34574e9-d367-4ec1-a84d-a2e65e4fea05”
},
{
“x”: 424.5,
“y”: 520.0,
“width”: 97.0,
“height”: 80.0,
“confidence”: 0.8086674213409424,
“class”: “trees”,
“class_id”: 0,
“detection_id”: “096f6e4f-5cf5-451a-a0d1-2d1352eeba21”
},
{
“x”: 497.5,
“y”: 557.0,
“width”: 83.0,
“height”: 66.0,
“confidence”: 0.8046267032623291,
“class”: “trees”,
“class_id”: 0,
“detection_id”: “bb309d7e-0560-4f96-b0f1-9e23c2666940”
},
{
“x”: 527.5,
“y”: 502.5,
“width”: 83.0,
“height”: 83.0,
“confidence”: 0.7917442917823792,
“class”: “trees”,
“class_id”: 0,
“detection_id”: “5c9dbcc4-8d2f-46ee-9294-923e89408c4e”
}
]
}
and this is the visualize model response {
“predictions”: [
{
“x”: 364.5,
“y”: 131.5,
“width”: 69,
“height”: 57,
“confidence”: 0.851,
“class”: “trees”,
“class_id”: 0,
“detection_id”: “68e75e5e-5d70-4a6d-bbc0-1d4b6bc8578d”
},
{
“x”: 412.5,
“y”: 68,
“width”: 61,
“height”: 60,
“confidence”: 0.839,
“class”: “trees”,
“class_id”: 0,
“detection_id”: “e3cd7571-7481-4ae9-bec0-999f0b4e652c”
},
{
“x”: 359,
“y”: 221.5,
“width”: 94,
“height”: 99,
“confidence”: 0.836,
“class”: “trees”,
“class_id”: 0,
“detection_id”: “67a52ec2-a273-4e06-9324-7fb39f5e824d”
},
{
“x”: 330.5,
“y”: 456.5,
“width”: 95,
“height”: 91,
“confidence”: 0.836,
“class”: “trees”,
“class_id”: 0,
“detection_id”: “1bc8bc3f-63cd-47ba-915e-5bf8d9aea0bc”
},
{
“x”: 703,
“y”: 356,
“width”: 186,
“height”: 222,
“confidence”: 0.823,
“class”: “trees”,
“class_id”: 0,
“detection_id”: “4075be3d-c417-4198-a5c2-950a5c73628d”
},
{
“x”: 423,
“y”: 519.5,
“width”: 98,
“height”: 81,
“confidence”: 0.811,
“class”: “trees”,
“class_id”: 0,
“detection_id”: “e8f36abe-7d2d-4410-aef4-c44804b218ec”
},
{
“x”: 497,
“y”: 557,
“width”: 84,
“height”: 68,
“confidence”: 0.805,
“class”: “trees”,
“class_id”: 0,
“detection_id”: “d371e35a-8255-4e8e-8f4d-8c20ef647842”
},
{
“x”: 528,
“y”: 502.5,
“width”: 84,
“height”: 83,
“confidence”: 0.785,
“class”: “trees”,
“class_id”: 0,
“detection_id”: “b116a571-8885-4fb1-b89d-423d0d6e0455”
}
]
} and this is the image so basically i want to detect trees from siteplan images
