def scatter(ss, fig):
db = {}
for im_file in nomes:
nmbe = desc.bendenergy(path + im_file, ss)
db[im_file] = numpy.hstack((cl[im_file], numpy.log(nmbe())))
# nome das figuras
data1 = numpy.array([db[i] for i in db.keys()])
Y = data1[:, 0].astype(int)
X1 = scale(data1[:, 1:])
#iso = Isomap(n_neighbors=98, max_iter= 2500)
mds = MDS(n_init=20, dissimilarity='euclidean', max_iter=2500)
#X1 = iso.fit_transform(data1[:,1:])
X1 = mds.fit_transform(data1[:, 1:])
#r = ((pdist(data1[:,1:]) - pdist(X1))**2).sum()
#s = ((pdist(X1)-pdist(X1).mean())**2).sum()
#R2 = 1-r/s
#print R2
data = numpy.vstack((Y, X1.transpose())).transpose()
db = dict(zip(db.keys(), data))
ax = PLT.subplot(111)
PLT.gray()
PLT.xlim((-5., 5.))
PLT.ylim((-5., 5.))
for im in db.keys():
# add a first image
img = Image.open(path + im)
img.thumbnail((100, 100), Image.ANTIALIAS)
#img = PIL.ImageOps.invert(img.convert("L"))
img = img.convert("RGBA")
datas = img.getdata()
newData = []
for item in datas:
if item[0] == 255 and item[1] == 255 and item[2] == 255:
newData.append((255, 255, 255, 0))
else:
newData.append(ImageColor.getrgb(colors[int(db[im][0])]))
img.putdata(newData)
imagebox = OffsetImage(numpy.array(img), zoom=.15)
xy = [db[im][1], db[im][2]] # coordinates to position this image
ab = AnnotationBbox(imagebox,
xy,
xybox=(2.5, -2.5),
xycoords='data',
boxcoords="offset points",
frameon=False)
ax.add_artist(ab)
ax.grid(False)
#!/usr/bin/python
import sys
import descritores
import pylab
import cPickle
import metrics
db = cPickle.load(open(sys.argv[1] + "/classes.txt"))
#names = cPickle.load(open(sys.argv[1]+"/names.pkl"))
names = db.keys()
scales = pylab.loadtxt(sys.argv[2])
X = [
pylab.vstack(([db[f] for f in names],
pylab.array([
pylab.log(descritores.bendenergy(sys.argv[1] + f, s)())
for f in names
]).T)).T for s in scales
]
for x in X:
s = metrics.silhouette(x[:, 1:], x[:, 0].astype(int) - 1)
print pylab.mean(s), pylab.std(s)
for n, x in zip(['nmbe_pso.pkl', 'nmbe_de.pkl', 'nmbe_sa.pkl'], X):
with open(n, "wb") as f:
cPickle.dump(dict(zip(names, x)), f)
def ff(cc,s): return np.log(desc.bendenergy(cc,s)())
cl = pickle.load(f)
f.close()
db = {}
# Definição as escalas de acordo com Costa (1996)
S = 16
tau_max = 128.
tau_min = 0.8
oct_l = scipy.array([math.sqrt(2)**l for l in range(S)])
sigma = tau_min + (tau_max - tau_min) * (oct_l - math.sqrt(2)) / (oct_l.max() -
math.sqrt(2))
print(sigma)
#sigma= scipy.logspace(0.1,1.65,N)
tt = time.time()
print("feature extraction")
for im_file in cl.keys():
print(im_file)
nmbe = descritores.bendenergy(diretorio + im_file, sigma)
db[im_file] = scipy.hstack((cl[im_file], scipy.log(nmbe())))
with open(sys.argv[2], 'wb') as f:
pickle.dump(db, f)
print("done feature extraction in {0} seconds".format(time.time() - tt))
def ff(cc, s):
return np.log(desc.bendenergy(cc, s)())
if caminho:
walk(caminho, visit, lista_de_arquivos)
else:
os.exit(-1)
if passo == "0":
# Le arquivo de entrada e gera descritores
# plota graficos e gera arquivos de saida
sigma = np.logspace(sigma_min, sigma_max, sigma_n)
figure(1)
for im_file, out_file, plt_file in zip(lista_de_arquivos[0], lista_de_arquivos[1], lista_de_arquivos[2]):
# print im_file,"\t",out_file,"\t",plt_file,"\n"
output = cStringIO.StringIO()
fout = open(out_file, "w")
k = curvatura(im_file, sigma)
nmbe = bendenergy(k)
for a in nmbe.phi[::-1]:
output.write("{0: < 5,.3f} ".format(math.log(a)))
output.write("\n")
fout.write(output.getvalue())
output.close()
fout.close()
a = np.loadtxt(out_file)
plot(np.log(1 / sigma[::-1]), a, ".")
savefig(plt_file)
clf()
elif passo == "1":
# dicionario que associa cada figura a classe que esta pertence
classe_dic = dict(
#!/usr/bin/python import sys import descritores import pylab import cPickle import metrics db = cPickle.load(open(sys.argv[1]+"/classes.txt")) #names = cPickle.load(open(sys.argv[1]+"/names.pkl")) names = db.keys() scales = pylab.loadtxt(sys.argv[2]) X = [pylab.vstack(([db[f] for f in names],pylab.array([pylab.log(descritores.bendenergy(sys.argv[1]+f,s)()) for f in names]).T)).T for s in scales] for x in X: s = metrics.silhouette(x[:,1:],x[:,0].astype(int)-1) print pylab.mean(s),pylab.std(s) for n,x in zip(['nmbe_pso.pkl','nmbe_de.pkl','nmbe_sa.pkl'],X): with open(n,"wb") as f: cPickle.dump(dict(zip(names,x)),f)
8:"#aa2a2a",9:"#d4552a",10:"#ff7f2a",11:"#00d455",12:"#2aff55",13:"#7f2a55",14:"#aa5555",15:"#d47f55",
16:"#ffaa55",17:"#2a2a00",18:"#2aff7f",19:"#7f007f",20:"#aa557f",21:"#d47f7f",22:"#ffaa7f",23:"#00ffaa",
24:"#5500aa",25:"#7f2aaa",26:"#aa55aa",27:"#d4aaaa",28:"#ffd4aa",29:"#2a00d4",30:"#552ad4",31:"#7f55d4",
32:"#aa7fd4"}
path = sys.argv[1]
with open(path+"classes.txt","r") as f:
with open(path+"names.pkl","r") as g:
cl = cPickle.load(f)
nomes = cPickle.load(g)
db = {}
for im_file in nomes:
nmbe = desc.bendenergy(path+im_file,sigma)
db[im_file] = numpy.hstack((cl[im_file],numpy.log(nmbe())))
# nome das figuras
data1 = numpy.array([db[i] for i in db.keys()])
Y = data1[:,0].astype(int)
X1 = scale(data1[:,1:])
s = silhouette.silhouette(X1,Y-1)
print numpy.median(numpy.abs(1.- s))
#iso = Isomap(n_neighbors=98, max_iter= 2500)
mds = MDS(n_init = 20,dissimilarity = 'euclidean',max_iter = 2500)
#X1 = iso.fit_transform(data1[:,1:])
X1 = mds.fit_transform(data1[:,1:])
warnings.simplefilter("ignore")
ss = pylab.loadtxt(sys.argv[1])
path = sys.argv[2]
dim = ss.shape[1]-8
with open(path+"classes.txt","rb") as f:
with open(path+"names.pkl","rb") as g:
cl = pickle.load(f)
nomes = pickle.load(g)
clf = neighbors.KNeighborsClassifier(n_neighbors = 3)
it = model_selection.RepeatedStratifiedKFold(n_splits = 5,n_repeats = 50)
for s in ss:
sigma = s[4:4+dim]
SI,DB,CH = s[dim+4],s[dim+5],s[dim+6]
db = {}
for im_file in nomes:
nmbe = desc.bendenergy(path+im_file,sigma)
db[im_file] = np.hstack((cl[im_file],np.log(nmbe())))
# nome das figuras
Y = np.array([db[i][0] for i in db.keys()]).astype(int)
X = scale(np.array([db[i][1:] for i in db.keys()]))
res = model_selection.cross_val_score(clf,X,Y,cv = it,scoring = "accuracy")
st = str("{0} {1} {2} {3} {4} {5}").format(s[1],s[2],s[3],SI,DB,CH)
print(" ".join([st]+["{:2.2f}".format(i) for i in sigma]+["{:0.2f} {:0.2f}".format(res.mean(),res.std())]))
print
