Skip to first unread message
Param Rajpura
May 25, 2015, 8:16:43 AM5/25/15
to caffe...@googlegroups.com
Hi !!!
I am working on converting a finetuned AlexNet model to fully-convolutional model for obtaining a prediction map.
While referring the Net surgery, i following highlighted statement in the paragraph was difficult to visualize......
To do so we translate the InnerProduct matrix multiplication layers of CaffeNet into Convolutional layers. This is the only change: the other layer types are agnostic to spatial size. Convolution is translation-invariant, activations are elementwise operations, and so on. The fc6 inner product when carried out as convolution by fc6-conv turns into a 6 \times 6 filter with stride 1 on pool5. Back in image space this gives a classification for each 227 × 227 box with stride 32 in pixels. Remember the equation for output map / receptive field size, output = (input - kernel_size) / stride + 1, and work out the indexing details for a clear understanding.
I have tried to workout and experiment with appropriate input sizes like 475*475 for a 10 * 10 prediction map...or
835*835 for a 20*20 prediction map .....the 227x227 box size and 32 stride fit in but why does this happen????
1. We had trained the model with 227*227 images is that the reason of box size and from where does this stride value 32 come in???
2. Does it have to do anything with the property of the fully-connected layers we converted to Convolutional layers????
Need serious help in understanding the concepts!!!!
Already had a couple of sleepless nights!!!!
I am working on converting a finetuned AlexNet model to fully-convolutional model for obtaining a prediction map.
While referring the Net surgery, i following highlighted statement in the paragraph was difficult to visualize......
To do so we translate the InnerProduct matrix multiplication layers of CaffeNet into Convolutional layers. This is the only change: the other layer types are agnostic to spatial size. Convolution is translation-invariant, activations are elementwise operations, and so on. The fc6 inner product when carried out as convolution by fc6-conv turns into a 6 \times 6 filter with stride 1 on pool5. Back in image space this gives a classification for each 227 × 227 box with stride 32 in pixels. Remember the equation for output map / receptive field size, output = (input - kernel_size) / stride + 1, and work out the indexing details for a clear understanding.
I have tried to workout and experiment with appropriate input sizes like 475*475 for a 10 * 10 prediction map...or
835*835 for a 20*20 prediction map .....the 227x227 box size and 32 stride fit in but why does this happen????
1. We had trained the model with 227*227 images is that the reason of box size and from where does this stride value 32 come in???
2. Does it have to do anything with the property of the fully-connected layers we converted to Convolutional layers????
Need serious help in understanding the concepts!!!!
Already had a couple of sleepless nights!!!!
saeed masoomi
Feb 13, 2018, 11:42:11 AM2/13/18
to Caffe Users
Hi, I'm stuck in stride concept? do you get any idea why stride is 32?
Reply all
Reply to author
Forward
0 new messages