def __init__(self, mode='train', roiSize=224,
balance=False, transform=None):
super(SmallPatchDataset).__init__()
self.roiSize = roiSize
self.transform = transform
self.patch_dir = os.path.expanduser("~/patch%s" % self.roiSize)
self.gene_label = datautil.load_gene_label(0)
all_genes = datautil.get_gene_list(0)
all_genes = [gene for gene in all_genes if gene in self.gene_label]
spivot = int(len(all_genes) * 0.7)
epivot = int(len(all_genes) * 0.9)
if mode == "train":
self.genes = all_genes[:spivot]
if balance:
self.genes = datautil.get_balanced_gene_list(self.genes, 0)
elif mode == "val":
self.genes = all_genes[spivot:epivot]
elif mode == "test":
self.genes = all_genes[epivot:]
else:
raise Exception("Unknown mode in SPatchAllDataset", mode)
with open("roi/roi%d.json" % self.roiSize, 'r') as f:
self.roi = json.load(f)
self.gene_img_pids = [(gene, img, pid) for gene in self.genes
for img in self.roi[gene].keys()
for pid in range(len(self.roi[gene][img]))]
if mode == "train":
idx = np.random.permutation(range(len(self.gene_img_pids)))
self.gene_img_pids = [self.gene_img_pids[i] for i in idx]
def load_val_data(size=1, balance=False, fv='res18-128'):
gene_list = datautil.get_val_gene_list(size)
if balance:
gene_list = datautil.get_balanced_gene_list(gene_list, size)
return _load_data(gene_list, size=size, fv=fv)
def load_train_data(size=0, balance=False, fv='matlab'):
gene_list = datautil.get_train_gene_list(size)
if balance:
gene_list = datautil.get_balanced_gene_list(gene_list, size)
return _load_data(gene_list, size=size)